2011-10-23T00:02:00 janzert : should we expect attackradius2, viewradius2 and spawnradius2 to change in the course of the tournamnet, or will they be fixed to their current values throughout ? 2011-10-23T00:02:26 Has anyone figured out what spawradius is for? 2011-10-23T00:02:56 chances are they won't change but if they do your bot should be able to handle it 2011-10-23T00:03:21 SharkMonkey: that is the distance an ant will collect food 2011-10-23T00:04:55 janzert, Aha. Thanks 2011-10-23T00:04:57 I can't get clojure to work locally but I if I submit it, it passes tests. Is there anywhere I can get more details on how the server compiles and runs it so i can replicate that environment? 2011-10-23T00:05:00 < Alexer> jmcarthur, but yeah, i've no idea if you should trust that or the visualization :P 2011-10-23T00:05:36 jmcarthur, looking at http://aichallenge.org/game_settings.php apparently the servers use viewradius2 of 77 2011-10-23T00:05:50 *** Axoren has joined #aichallenge 2011-10-23T00:05:50 are AbstractSystemInputParser and AbstractSystemInputReader useful? 2011-10-23T00:06:07 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T00:06:14 I'm having errors with those, actually. Java Starter is giving me compiling issues 2011-10-23T00:06:25 so it's understandable that our formulas to get the pixel count for viewradius2 of 55 would differ from the pixel count observed in the replays :P 2011-10-23T00:06:29 not compiling. test execution* 2011-10-23T00:06:53 *** OrangeXP has left #aichallenge 2011-10-23T00:07:07 *** TheOnly92 has joined #aichallenge 2011-10-23T00:07:15 cutoff: they are useful reminders of why I hate java so much 2011-10-23T00:08:19 Anyone here use the java starter files successfully? 2011-10-23T00:08:48 If the java ones are broken that might explain why the clojure ones are having problems... 2011-10-23T00:08:49 Axoren: yes, are you having issues? 2011-10-23T00:09:10 Specifically after uploading and tests are run. 2011-10-23T00:09:15 (minus the fact the server doesn't seem to want to play matches at the moment) 2011-10-23T00:09:21 I'm getting null pointer exceptions in the main method provided 2011-10-23T00:09:51 odd 2011-10-23T00:09:53 I'll PM you the error message. 2011-10-23T00:10:16 I really hope that went to you. 2011-10-23T00:10:27 Haven't used IRC since I was in high school. 2011-10-23T00:11:04 janzert: the main issue is the query to get the match qualities goes from 0.01 seconds to 20-40 seconds when I join with the tmp_games (games in last 24 hours), so I'm going to try and rewrite it without the join 2011-10-23T00:11:44 janzert: did you ever bring up the old beta to check the server_info.php settings? 2011-10-23T00:13:08 *** avers_ has joined #aichallenge 2011-10-23T00:13:09 McLeopold, missing an index somewhere? 2011-10-23T00:14:50 When a bunch of ants are backed up in the hive, is that information explicitly given to the user? 2011-10-23T00:15:53 *** hmhm has joined #aichallenge 2011-10-23T00:16:49 can C++ programs depend on boost? 2011-10-23T00:17:01 *** jtamer has quit IRC (Ping timeout: 260 seconds) 2011-10-23T00:17:22 You can always bundle boost with your pogram if it comes to it 2011-10-23T00:17:24 how do I pass my bot to the test_bot? 2011-10-23T00:17:51 *** levis501 has quit IRC (Ping timeout: 256 seconds) 2011-10-23T00:18:00 SharkMonkey: re hive count, no 2011-10-23T00:18:31 iemand, ./test_bot.sh "command to launch my bot" 2011-10-23T00:18:51 ChrisH_, Didn't think so. Thanks. 2011-10-23T00:18:52 rofer: how so? timeout? 2011-10-23T00:19:09 McLeopold: the only difference is the one I thought might be, cutoff_turn 2011-10-23T00:19:09 SharkMonkey: yeah, but that seems like kind of a waste. 2011-10-23T00:19:32 boost fits in 2MB? :X 2011-10-23T00:19:33 otzi, There is this tool you can use to just extract the bits of boost you need. 2011-10-23T00:19:50 hm 2011-10-23T00:20:01 Which can really cut the size down. Assuming you don't use things with a million dependencies 2011-10-23T00:20:15 *** dom7b5 has joined #aichallenge 2011-10-23T00:20:32 and I think I'd rather leave it at 100 for now until we are seeing an appreciable percentage of problems 2011-10-23T00:21:18 do you know the tool's name? 2011-10-23T00:21:30 *** TheOnly92 has quit IRC (Ping timeout: 265 seconds) 2011-10-23T00:22:09 janzert: okay, then we should update the tools and game engine to default to radius 77 2011-10-23T00:23:09 otzi, Had to ask in #boost, to remember :) It's called bcp 2011-10-23T00:23:21 cool, thanks :) 2011-10-23T00:23:31 Comes bundled with boost 2011-10-23T00:23:43 McLeopold: they do don't they? 2011-10-23T00:23:49 at least ants.py does 2011-10-23T00:24:26 ahh, I guess not 2011-10-23T00:25:36 great 2011-10-23T00:26:24 aichallenge: janzert * r2666c0b / ants/playgame.py : Set default viewradius2 to match server setting - http://git.io/dMVCtw 2011-10-23T00:26:59 there now they do :) 2011-10-23T00:27:36 *** murilo has joined #aichallenge 2011-10-23T00:28:13 that means my tutorial map won't work the same :( 2011-10-23T00:28:53 :( 2011-10-23T00:32:24 janzert, btw, have you considered running an official tcp server? 2011-10-23T00:33:14 Anyone want to wish this nub some luck on his first successful upload? 2011-10-23T00:33:34 Axoren, Break a leg 2011-10-23T00:34:14 Alexer: we generally have our hands full keeping the current official servers going, so considered but rejected for now at least 2011-10-23T00:34:16 That doesn't work so well for coders as it does performers 2011-10-23T00:34:24 That saying sounds rather threatening coming from a SharkMonkey, 2011-10-23T00:34:29 janzert, ok 2011-10-23T00:34:37 xD 2011-10-23T00:35:21 Axoren, It does seem kind of rude, when said to a stranger over irc. I apologize 2011-10-23T00:35:23 *** delt0r_ has quit IRC (Ping timeout: 245 seconds) 2011-10-23T00:35:26 *** twymer has joined #aichallenge 2011-10-23T00:35:40 The gesture was appreciated! 2011-10-23T00:35:43 Don't fret! 2011-10-23T00:37:56 *** UltraMegaSupreme has quit IRC (Ping timeout: 265 seconds) 2011-10-23T00:39:52 *** FranPeruzzi has quit IRC (Ping timeout: 265 seconds) 2011-10-23T00:40:35 is there something wrong with the compiler? 2011-10-23T00:40:42 this is what i submitted: http://imageshack.us/photo/my-images/198/screenshot20111022at103.png/ 2011-10-23T00:41:29 what error did you get? 2011-10-23T00:41:38 this is the error i'm getting: http://imageshack.us/photo/my-images/405/screenshot20111022at103.png/ 2011-10-23T00:41:49 MyBot.* is not found 2011-10-23T00:41:54 compiles fine and it runs 2011-10-23T00:42:31 iemand: are you on a mac? 2011-10-23T00:42:40 yeah 2011-10-23T00:43:13 default mac zip creates hidden directories that stop the submission from unpacking correctly 2011-10-23T00:43:44 tar packs zip archives, doesn't it? 2011-10-23T00:43:54 isn't tar part of UNIX? 2011-10-23T00:43:58 can he use that? 2011-10-23T00:44:10 I think tar doesn't pack zip, it's a different format 2011-10-23T00:44:18 the contest will accept tar files 2011-10-23T00:44:26 yeah tar is standard 2011-10-23T00:44:30 I'll try that, thaks 2011-10-23T00:44:32 thanks* 2011-10-23T00:44:46 *** levis501 has joined #aichallenge 2011-10-23T00:44:56 specifically tar.gz or .tgz extensions 2011-10-23T00:45:17 *** Purrp has quit IRC (Quit: Page closed) 2011-10-23T00:45:18 ... I can submit files while at work now that I know this... 2011-10-23T00:45:43 you can play at work now :P 2011-10-23T00:45:55 oh, and .tar.xz, tar.bz2, .txz, .tbz apparently as well 2011-10-23T00:46:33 Where will I find my bot's recently participated games/when my bot will participate in a game? 2011-10-23T00:47:12 also if someone can get me a macos zipped file I could try and make the worker correctly ignore the hidden directories 2011-10-23T00:47:29 Axoren: on your profile page 2011-10-23T00:47:37 janzert: which hidden directories are you talking about? 2011-10-23T00:47:45 The games will be listed along with my submission history? 2011-10-23T00:47:55 janzert: .DS_Store i imagine? 2011-10-23T00:48:05 I know macos adds them by default to a zip file but I don't know offhand what they are called 2011-10-23T00:48:05 yes, right after your personal info and before submissions 2011-10-23T00:48:14 that's why I need an example zip :) 2011-10-23T00:48:19 Thanks, bro. 2011-10-23T00:48:19 look at anyone else's profile for an idea 2011-10-23T00:48:33 my submissions are all from osx using the commandline zip tool which doesn't do that 2011-10-23T00:48:48 *** delt0r_ has joined #aichallenge 2011-10-23T00:48:59 *** goffrie has quit IRC (Remote host closed the connection) 2011-10-23T00:49:09 yeah, my macos zips are using ant, which doesn't do it either 2011-10-23T00:49:16 i can make you one, though, janzert 2011-10-23T00:49:22 a1k0n: ahh, so there must be a gui zip on osx? 2011-10-23T00:49:34 similiar to windows I imagine 2011-10-23T00:49:52 yeah but it doesn't include anything special 2011-10-23T00:49:58 I see stats for players ahead and time before next game, but games aren't being played on my bot 2011-10-23T00:50:04 i just tried it 2011-10-23T00:50:10 I've received a lot of zips with .DS_Store over the years. Now I know how they were generated :P 2011-10-23T00:50:13 tar -zcvf submission.22.10.tar.gz submission.22.10 2011-10-23T00:50:17 are osx zips reported to be broken? 2011-10-23T00:50:21 still giving the same issue 2011-10-23T00:50:30 all I know is we tracked it down to that in the beta 2011-10-23T00:51:03 it seems like it actually had more than one extra directory 2011-10-23T00:51:04 maybe just add 'find . -name .DS_Store -exec rm -rf {} \;' to the script? 2011-10-23T00:51:15 oh 2011-10-23T00:51:18 resource forks 2011-10-23T00:51:24 i dunno if xcode adds some special mojo 2011-10-23T00:51:35 but i bet zip does something goofy to support them 2011-10-23T00:51:46 *** stan_ has joined #aichallenge 2011-10-23T00:51:46 I'll boot up my pc and zip it on there 2011-10-23T00:51:58 hmmm 2011-10-23T00:51:59 weird 2011-10-23T00:52:08 iemand: what is your username in the contest? 2011-10-23T00:52:11 I created a zip with finder, but everything seems legit on win7: 2011-10-23T00:52:12 http://www.mediafire.com/?xq12191pm27w42a 2011-10-23T00:52:15 yeah, so I think the server's skipping my bot's turn or something... 2011-10-23T00:52:21 I'll try taking a look at the submission file 2011-10-23T00:52:32 oh, Dominater 2011-10-23T00:52:36 *** stan_ has quit IRC (Client Quit) 2011-10-23T00:52:37 forgot I put in a different name 2011-10-23T00:52:40 Axoren: there is a pretty big backlog of games right now, so it might be a while, be patient, the devs are working on speeding it up. 2011-10-23T00:53:01 *** lavalamp has quit IRC (Quit: Page closed) 2011-10-23T00:53:11 So the stats it's telling me are inaccurate. 2011-10-23T00:53:39 More or less 2011-10-23T00:53:40 Axoren: what is it telling you? 2011-10-23T00:53:46 what is the variable spawnradius2 used for? 2011-10-23T00:53:51 next game in 0 minutes, 0 people ahead 2011-10-23T00:54:00 then it changes to different numbers and climbs back down 2011-10-23T00:54:21 *** stano has joined #aichallenge 2011-10-23T00:54:35 *** J0KRESTER has joined #aichallenge 2011-10-23T00:54:38 iemand: http://junecloud.com/software/mac/junecloud-automator-actions.html?site=junecloud&offset=-1 2011-10-23T00:54:48 *** amstan has joined #aichallenge 2011-10-23T00:54:48 *** ChanServ sets mode: +o amstan 2011-10-23T00:54:48 Axoren: If this is your first ever submission, then you are going to be near the front of the queue, but because the servers were down or very slow for most of today, there are a lot of others in your same position. 2011-10-23T00:55:15 "create clean archive" creates a zip archive without mac folder configurations 2011-10-23T00:55:17 I'm not begging for a speed up. 2011-10-23T00:55:23 I'm just curious 2011-10-23T00:55:28 Axoren: I have hear there may be a bug with that counter as well, so take it with a grain of salt. 2011-10-23T00:55:37 heard 2011-10-23T00:55:37 Alright 2011-10-23T00:56:02 That just means sleep instead of waiting to see the field test before bed. 2011-10-23T00:56:06 Thanks 2011-10-23T00:56:13 *** kincody has joined #aichallenge 2011-10-23T00:56:52 Axoren: no problem 2011-10-23T00:57:15 And by sleep, I mean reddit @_@ 2011-10-23T00:58:15 *** stano has quit IRC (Remote host closed the connection) 2011-10-23T00:58:59 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T00:59:19 *** Pawel__ has joined #aichallenge 2011-10-23T00:59:20 *** stano has joined #aichallenge 2011-10-23T00:59:39 Hello 2011-10-23T00:59:59 *** yoden has quit IRC (Quit: Leaving.) 2011-10-23T01:01:22 iemand: hmm, as suspected the zip you uploaded has an extra directory (__MACOSX) but I'm not sure why the tar file didn't work 2011-10-23T01:01:43 Does anyone know why there is no documentation for that contest????? Such as some api (like) documentation for all the methods that we can use writting our programs??? Like this we have to keep guessing what each method and class do... It is unecesary. 2011-10-23T01:02:05 ????? 2011-10-23T01:02:13 Pawel__: you have to relax 2011-10-23T01:02:15 it added a .ds_store to the tar.gz as well 2011-10-23T01:02:16 Pawel__: the game specification has the api 2011-10-23T01:02:17 a submission.22.10/.DS_Store 2011-10-23T01:02:33 Pawel__, Also, the python starter kit is well commented and straightforward 2011-10-23T01:02:42 iemand: but since it is under the submission directory it shouldn't have hurt 2011-10-23T01:02:56 Pawel__: what language are you even talking about? 2011-10-23T01:02:58 Next game should be within -0.1 minutes. 2011-10-23T01:03:00 Oo 2011-10-23T01:03:05 Where can u find it - i saw only brief description and no docs - ia m using java 2011-10-23T01:03:22 http://ai-contest.com/specification.php 2011-10-23T01:03:27 Pawel__, The api is just text in and text out. 2011-10-23T01:03:30 that's what your program has to follow 2011-10-23T01:04:07 oh ffs, Compile Error: error while compiling submission 2011-10-23T01:04:11 you could write it from ground up, or use the starter packages to help you out 2011-10-23T01:04:14 zipped and uploaded with a pc 2011-10-23T01:04:22 Pawel__: all bot communication happens on stdio so there is no direct method or function calls 2011-10-23T01:04:34 exact same error 2011-10-23T01:05:12 janzert: do you know how i can make sure a github wiki page makes it onto the specs page on the website? (i want Ants food harvesting to be between cutoff rules and ant spawning) 2011-10-23T01:05:30 Pawel__, you remind me of Mr. Chekov. :P 2011-10-23T01:05:41 antimatroid: look at specification.php in the repo 2011-10-23T01:05:44 *** avers_ has joined #aichallenge 2011-10-23T01:05:45 thanks. So I don't actually call any methods from main function??? Is that what u mean - sorry if this is a silly question. 2011-10-23T01:05:49 Why??? 2011-10-23T01:05:55 basically it has special comments in the source to insert the wiki pages 2011-10-23T01:06:07 I'm having a lot of trouble avoiding timeout doing GA per ant 2011-10-23T01:06:09 I've read it and I'm using C++ starter package. It has been very easy to follow. I don't know how the Java package is, but I think any package will be small enough that it should be easy to follow. 2011-10-23T01:06:15 why is he using so many question marks? 2011-10-23T01:06:46 sorry 2011-10-23T01:06:49 After my status is "ready to play" should I do something to play? 2011-10-23T01:07:16 murilo: nope, the server does all the work of making matchups and playing the games 2011-10-23T01:07:17 Pawel__: basically you can write the bot however you want, all the interfacing with the engine is done by printing to stdout and reading from stdin 2011-10-23T01:07:30 janzert: it should just work if i add in the single extra line to that file? 2011-10-23T01:07:42 murilo: but it is running slow so be patient, the devs are working on getting it to go faster 2011-10-23T01:07:47 antimatroid: so far as I know yes 2011-10-23T01:07:50 *** espes has quit IRC (Ping timeout: 276 seconds) 2011-10-23T01:07:54 okay 2011-10-23T01:08:02 *** xunie-laptop has quit IRC (Quit: Leaving) 2011-10-23T01:08:10 ChrisH_: ok, thanks! 2011-10-23T01:08:47 *** Box has joined #aichallenge 2011-10-23T01:09:05 *** Pawel__ has quit IRC (Quit: Page closed) 2011-10-23T01:09:13 *** Box is now known as Guest29562 2011-10-23T01:09:34 only the .java files have to be uploaded right? 2011-10-23T01:09:43 *** Guest29562 is now known as Boxx 2011-10-23T01:10:03 janzert: take a look at http://aichallenge.org/profile.php?user=49 -- he's really making a lot of submissions and being rewarded with lots of games for some reason 2011-10-23T01:12:15 aichallenge: Nick Ham * rcbeb386 / website/specification.php : Added a section to the specs page on food harvesting - http://git.io/qaVxdw 2011-10-23T01:12:44 *** kincody has quit IRC (Ping timeout: 265 seconds) 2011-10-23T01:12:45 we should shame people by letting everyone sort the leaderboard by number of submissions 2011-10-23T01:13:16 McLeopold: see ChrisH_ comment above ^^ 2011-10-23T01:13:18 antimatroid: it already has version count 2011-10-23T01:13:26 ChrisH_: he doesn't have lots of games, he has 1 game from his latest version 2011-10-23T01:13:42 *** bobbydroptable has joined #aichallenge 2011-10-23T01:13:55 ah... do ai/mlclass homework, or work on bot... 2011-10-23T01:13:57 decisions decisions 2011-10-23T01:14:00 amstan: he has more games than most anyone else 2011-10-23T01:14:07 basically one per submission 2011-10-23T01:14:18 bobbydroptable: what's better than actual experience? :P 2011-10-23T01:14:33 amstan: look at the recent games log, he is in about every other or every third game 2011-10-23T01:14:38 hehehe exactly! 2011-10-23T01:14:54 he's gotten 9 or 10 games since the worker came back up 2011-10-23T01:14:56 ChrisH_: well, how would you go about preventing this? 2011-10-23T01:15:15 amstan: he's not supposed to be bumped to the front of the line on submission 2011-10-23T01:15:24 the manager's doing exactly what he's supposed to be doing, give games for the gameless 2011-10-23T01:15:30 if he resubmits, he's gameless 2011-10-23T01:15:31 no one is 2011-10-23T01:15:32 amstan: throttle games on a per player level rather than per submission 2011-10-23T01:15:43 no, it's suppose to go by the player's games not the submissions 2011-10-23T01:15:45 server back up? darn, means my score will drop soon lol 2011-10-23T01:15:49 it is per user, not per submission 2011-10-23T01:15:58 McLeopold: oh, then it's broken 2011-10-23T01:16:13 every time i submitted a new program, i've been immediately scheduled for a game 2011-10-23T01:16:14 could be, I'll look at it after the speed issue is resolved 2011-10-23T01:16:19 we could also send hims an email, to shame him from doing this 2011-10-23T01:16:23 till we fix it 2011-10-23T01:16:30 I wouldn't 2011-10-23T01:16:47 but, it resets my rank, so i don't want to do it as my bot gets better 2011-10-23T01:16:53 we did to that one guy last year that was resubmitting every minute 2011-10-23T01:16:54 ChrisH_: http://aichallenge.org/profile.php?user=49 again! :/ 2011-10-23T01:17:16 iemand: I found the problem 2011-10-23T01:17:22 awesome got mine to work 2011-10-23T01:17:31 I'm waiting so long 2011-10-23T01:17:40 that's a lot of versions 2011-10-23T01:17:42 you can't zip a directory with the files in it, you have to select the files and send those to a zipped folder 2011-10-23T01:17:49 maybe there could be a scoring system that penalises high number of submissions? 2011-10-23T01:17:52 going to see if I can fix osx zips as well 2011-10-23T01:17:54 McLeopold: are you happy with harvesting as the word for food being removed and placed into an ants hive? also have you considered making spawn2 = 0 now that ants spawn on hills? I think that'd mean food is never "destroyed"? 2011-10-23T01:18:14 *** Saulzar has joined #aichallenge 2011-10-23T01:18:33 antimatroid: I'll have to read it later 2011-10-23T01:18:52 think about spawn2 = 0, i think that's a pretty good idea 2011-10-23T01:19:43 antimatroid: that would also mean that ants would need to be allowed to move onto food, which I've heard some people say they are not allowed to do now. 2011-10-23T01:19:46 ChrisH_ what again? 2011-10-23T01:19:55 what did 49 do 'again'? 2011-10-23T01:20:06 oh, he's getting too many games 2011-10-23T01:20:14 although it does make collecting food slightly easier, as you don't have to worry about the possibility of some shortest paths collecting up to 3 food items simultaneously 2011-10-23T01:20:19 by making lots of submissions 2011-10-23T01:20:50 oh crap 2011-10-23T01:21:05 antimatroid: I don't care as far as my bots go, but the engine may need a tweek if spawnradius2 = 0, that's all I was saying 2011-10-23T01:21:21 McLeopold : I was also going to point something similar to what Chrish mentioned. ranking page looks bit odd this time. see everybody in top 10-20, has got around same number of games 2011-10-23T01:21:24 yeah i know what you meant, i was just making an observation :) 2011-10-23T01:22:15 aichallenge: McLeopold * r2b256bd / sql/2_generate_matchup.sql : fix to submission ordering - http://git.io/EeA9FA 2011-10-23T01:22:15 aichallenge: McLeopold * ra32f154 / (12 files in 8 dirs): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/12k-eQ 2011-10-23T01:22:48 McLeopold : although that looks fair, they have submited their versions at different times. so bit unusual how all get end up playing about same number of games. I think the scheduler is pushing recent submitters on top and then continue with them until they get about same number of games as others. 2011-10-23T01:22:49 ashoka: I think that's what we want, roughly equal distribution of games 2011-10-23T01:22:57 I'd like that :D 2011-10-23T01:23:16 janzert / amstan: specification.php is modified on the server? 2011-10-23T01:23:21 McLeopold : oh, you have already fixed :-) 2011-10-23T01:23:30 McLeopold: i didn't 2011-10-23T01:23:30 McLeopold: i did that, did i break it? 2011-10-23T01:23:33 so what does spawn2 = 0 do again? 2011-10-23T01:23:43 oh you mean spawnradius2 2011-10-23T01:23:46 ChrisH_: was totally right, I messed up and order by most game :( 2011-10-23T01:23:59 McLeopold: they're modded because of the setup script 2011-10-23T01:24:01 ouch 2011-10-23T01:24:06 lol ouch 2011-10-23T01:24:11 amstan: no, this is different I think 2011-10-23T01:24:13 McLeopold: feel free to git reset them, just ./setup.py when finished 2011-10-23T01:24:17 Chrish : it can't be. in this case there are people who have submitted 2 days ago, and people who submited just 2 hours ago. how can they both get the same number of total games after 2 hours ? :-) 2011-10-23T01:24:20 *** korfuri has joined #aichallenge 2011-10-23T01:24:27 McLeopold: how so? 2011-10-23T01:24:59 oh, this means the git repo changed, not the file was edited on the server 2011-10-23T01:24:59 bobbydroptable: yeah i mean make an ant reside on the food location after battle to harvest it 2011-10-23T01:25:06 Maybe limit number of games per time interval? X games maximum per Y minutes? 2011-10-23T01:25:07 ashoka: keep in mind that the servers have been very slow/down for much of today, so the current state of the system is not representative of steady state 2011-10-23T01:25:11 what does spawnradius2 mean? 2011-10-23T01:25:18 *** Guest__ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T01:25:32 spawn radius squared 2011-10-23T01:25:46 What's the spawn radius? I didn't mess with that stuff :\ 2011-10-23T01:25:49 yeah, but what is spawn radius? 2011-10-23T01:25:50 it should really be called something like harvestRadius now it was left from before hills 2011-10-23T01:25:56 Question of today: "What does spawnradius2 mean?" 2011-10-23T01:26:21 a relic from the beta apparently 2011-10-23T01:26:26 it is the radius an ant has to get within a food in order to spawn a new ant in the hive 2011-10-23T01:26:37 read tools/ants.py line 659 2011-10-23T01:26:39 OH! 2011-10-23T01:26:41 McLeopold: yes, that comes up whenever a file in the repo gets updated that is expanded on the server by setup.py 2011-10-23T01:26:41 theres a comment that describes it 2011-10-23T01:26:42 radius^2 really 2011-10-23T01:26:47 Chrish : Servers got slow due to pairing issue. probably it's the match making algorithm giving more priority to newer submitters, I don't know but top 20 odd playes who have submitted their versions days apart are unlikely to get the same number of games if the game rate is evenly distributed. 2011-10-23T01:26:52 we could change that to foodradius or gatherradius or harvestradius, but it might be too late 2011-10-23T01:27:07 That can make my bot 1 turn faster per farmer ant, effectively 2011-10-23T01:27:17 ChrisH_: so if the food is within the spawn radius the ant will eat it? I see. 2011-10-23T01:27:21 https://github.com/aichallenge/aichallenge/wiki/Ants-food-harvesting that explains the radius 2011-10-23T01:27:40 I've been programming them to use A* from their last unemployed position to the closest food's location, rather than adjacent to it. 2011-10-23T01:27:43 but how much does that actually impact ai , really... 2011-10-23T01:27:47 ashoka: there are two aspects to the scheduling, one is that every player gets an equal number of chances to be the "seed" player for any given game, then there is the method of chosing who else should be in that game. 2011-10-23T01:27:52 I think spawnradius2 has to stay but it could be undocumented with a better named paramater that means the same thing also sent 2011-10-23T01:27:53 i just tell ants to move towards food 2011-10-23T01:27:56 *** skunx has quit IRC (Quit: leaving) 2011-10-23T01:28:00 that gets a little ugly though 2011-10-23T01:28:13 does anyone's bot take spawnradius2 into account? 2011-10-23T01:28:16 ashoka: ideally the other players are in roughly the same skill range and haven't played each other recently... 2011-10-23T01:28:21 antimatroid: I've always used the term "gather", can we switch it? 2011-10-23T01:28:24 *** Boxx has quit IRC (Quit: Page closed) 2011-10-23T01:28:36 i use forage or gather 2011-10-23T01:28:40 I second that 2011-10-23T01:28:47 *** andy2011 has joined #aichallenge 2011-10-23T01:28:49 my behavioral agent is called GatheringAgent 2011-10-23T01:28:53 lol 2011-10-23T01:29:20 McLeopold: we can, i just thought harvest actually captured the notion of removing the food and placing it into the hive better, i think of gathering/collecting food as the whole process of moving to and picking up the food, but i don't mind 2011-10-23T01:29:41 ashoka: so the scheduling is a pretty dynamic system 2011-10-23T01:31:10 i'm also not sure of using exist and distinct, you might want to make that a bit less mathsy 2011-10-23T01:31:11 ashoka: but constraints were put into place to prevent players from getting too many games by not scheduling them (except as game seeds) if their # of games gets too far ahead of the rest of the field, and that constraint is what has leveled the numbers you are talking about 2011-10-23T01:32:04 *** iemand has quit IRC (Ping timeout: 265 seconds) 2011-10-23T01:32:46 bobbydroptable: yes, my bot uses spawnradius2 2011-10-23T01:33:12 bobbydroptable: some of the sample files during the beta had a value >1 for it 2011-10-23T01:33:20 *** thestinger has quit IRC (Quit: WeeChat 0.3.5) 2011-10-23T01:33:27 bobbydroptable: you should use all parameters, while it is desirable for both contestants and us that they wont change we do not guarantee it 2011-10-23T01:33:28 btw, the idea is to keep creating new games each year? Seems like last year it was a different one called PlanetWars. 2011-10-23T01:34:04 I believe so 2011-10-23T01:34:11 far more interesting that way 2011-10-23T01:34:19 in other words if you hard code parameters and we need to change it for whatever reason, you should not complain to us that your bot no longer works :P 2011-10-23T01:34:22 What does (>")> mean on the ranking list? 2011-10-23T01:34:42 it looks like kirby dancing 2011-10-23T01:34:52 otzi: planet wars and tron were both last year, ants was intended to be done earlier in the year but it took longer than we expected to make due to contributers time contraints 2011-10-23T01:34:59 Axoren: it's a bug 2011-10-23T01:35:10 just a question, but does the tutorial http://aichallenge.org/ants_tutorial.php instruction here work on a mac as well... I cant seem to get it to work without changing the instructions slightly.... 2011-10-23T01:35:18 it means that player just uploaded new code, but it hasn't been compiled yet 2011-10-23T01:35:27 *** hmhm has quit IRC (Ping timeout: 265 seconds) 2011-10-23T01:35:54 antimatroid: I see. Were they also this sophisticated with a pretty user interface and all that? 2011-10-23T01:36:01 Kirby shall compile the code with the power of dance 2011-10-23T01:36:21 no, the visualiser for this contest goes above and beyond those from planet wars and tron 2011-10-23T01:37:09 and so do a few other things 2011-10-23T01:37:19 I thought the planet wars visualizer was pretty. 2011-10-23T01:37:28 but the game was simpler 2011-10-23T01:37:31 yeh i know i meant that 2011-10-23T01:37:34 yeah it was cool, but i don't think it was as sophisticated as the ants one 2011-10-23T01:37:40 knowing this spawn radius stuff makes my bot so much faster 2011-10-23T01:37:42 spawnradius2 doesn't seem like a significant factor 2011-10-23T01:37:44 it has less to do 2011-10-23T01:38:05 i suppose you could say "stop pathfinding if distance less than radius"? 2011-10-23T01:38:07 bobbydroptable: you need to know which squares around food you need to find a shortest path to in order to harvest it 2011-10-23T01:38:07 I was amazed by it. I was expecting it to be just text input and output when I heard about it :P 2011-10-23T01:38:11 *** moai has quit IRC (Remote host closed the connection) 2011-10-23T01:38:33 antimatroid why can't you just search directly to the food ? 2011-10-23T01:38:33 I give up... 3:38AM 2011-10-23T01:38:44 if my ants are spending an extra turn gathering, they aren't actively attacking enemy ants 2011-10-23T01:38:46 if spawnradius2>= 2 then it's possibly to harvest across water/diagonals 2011-10-23T01:38:48 wow 2011-10-23T01:38:55 mmmm . 2011-10-23T01:38:59 that's a big deal 2011-10-23T01:39:12 it only takes one turn to raze a hill 2011-10-23T01:39:13 that is a good point , harvesting over water 2011-10-23T01:39:25 it is 3h40 indeed. I hate daylight saving time :( 2011-10-23T01:39:51 Axoren: if your bot moves within the radius of the food, the food will be gone, so it won't spend any extra turns? 2011-10-23T01:40:00 yes 2011-10-23T01:40:20 ah never mind *shrug* its my bot lol 2011-10-23T01:40:28 but if you spend an extra turn moving ONTO the tile that had the food, that's an issue 2011-10-23T01:40:51 otzi: same here... 2011-10-23T01:40:55 My bot does a lot more Search and Destroy 2011-10-23T01:41:10 ah you're talking about high level strategy, not algorithms 2011-10-23T01:41:30 but then an extra move would apply to everyone 2011-10-23T01:42:22 aichallenge: janzert * r4ef47a7 / worker/worker.py : Fix unpacking OSX zips and upload files with top directories - http://git.io/bgmbQw 2011-10-23T01:42:37 you mean an extra move would apply to everyone using an algorithm that assumed you needed to land on a food tile to take it 2011-10-23T01:42:37 Hmm - someone posted an improved haskell starter kit here https://github.com/qacek/ants-haskell/ 2011-10-23T01:42:39 *** twymer has quit IRC (Ping timeout: 248 seconds) 2011-10-23T01:42:51 some people cache tiles incase they lose sight of something 2011-10-23T01:43:02 "There was food here" 2011-10-23T01:43:29 ok, Mac zips should work now 2011-10-23T01:43:34 my algorithm just spends an extra turn saying "I already picked the food up, but I'm still going to move into that space. 2011-10-23T01:43:47 ah 2011-10-23T01:44:04 i compute every ant's behaviour on every move 2011-10-23T01:44:18 because i couldn't be stuffed trying to track every ant's state lol 2011-10-23T01:44:25 That leads to timeouts if you end up with too many ants 2011-10-23T01:44:31 so far it's worked ok :D 2011-10-23T01:44:53 i'm not sure how my modifications will fare, when the time comes 2011-10-23T01:45:09 Axoren: speak for yourself, I do the same and don't timeout 2011-10-23T01:45:31 I've seen people with really simple AI's timeout 2011-10-23T01:45:33 how's Go coming along? speedy? 2011-10-23T01:45:37 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T01:45:45 Axoren: simple AI, but slow code, yes 2011-10-23T01:45:57 or simple AI with bad infinite loop 2011-10-23T01:45:58 lol 2011-10-23T01:46:27 I just hope my A* algorithm works quickly 2011-10-23T01:46:33 anyone seen any map prediction going on based on the symmetry 2011-10-23T01:46:33 Axoren, i do that too. as a precaution i check how much time i have after each path calculation and bail out if it seems i might run out (thus some of my ants might stop on their tracks for a few turns if i ever came close to the timeout) 2011-10-23T01:46:53 Alexer: python starter kit does this 2011-10-23T01:46:58 Six: brilliant idae 2011-10-23T01:47:04 A* is easy to screw up and turn into an infinate loop 2011-10-23T01:47:11 Bot code can run very fast if you cache the results of certain computations that are always the same for a given board. 2011-10-23T01:47:25 Six: some guy was trying to do something like this , but he wanted to upload all the official maps 2011-10-23T01:47:34 Alexer, that's brilliant 2011-10-23T01:47:37 *** Palmik has joined #aichallenge 2011-10-23T01:48:02 If I ever get a situation where my bot ends up doing too much pathfinding in one turn, I'll have a huge issue 2011-10-23T01:48:21 my first bot only does a* within it's field of vision 2011-10-23T01:48:29 so i have this problem where my ants don't know what to do 2011-10-23T01:48:34 when there's nothing around them lol 2011-10-23T01:48:40 my 2nd bot is trying to solve that ==" 2011-10-23T01:48:40 Make them lemmings 2011-10-23T01:48:54 when mine have that issue, they weigh a tile by how close it is to other ants/hills 2011-10-23T01:48:54 at least lemmings walk in a straight line ==" 2011-10-23T01:48:56 and runs away 2011-10-23T01:49:06 Xenophobic ants, ftw 2011-10-23T01:49:18 how do they attack? 2011-10-23T01:49:19 lol 2011-10-23T01:49:28 Axoren, won't they all path to the same places then? 2011-10-23T01:49:50 No, I make sure that tiles already part of a "job" are weighted heavily 2011-10-23T01:50:31 they will also wait until an ant moves away from their target location 2011-10-23T01:50:42 giving the ant occupying that location time to run away like a xenophobe or acquire a new job 2011-10-23T01:51:10 i haven't yet decided what to do about the idle problem (rationally) 2011-10-23T01:51:13 Also, attacks only happen when food is not present and enemies are visible. 2011-10-23T01:51:22 If neither is present, xenophobia 2011-10-23T01:51:30 right so you run away from your own ants 2011-10-23T01:51:31 if none of the above are possible, random direction 2011-10-23T01:51:39 i had to do something, so currently they just use a longest-path algorithm and run in circles like headless chickens :D 2011-10-23T01:52:19 so they path to the other end of the area they see, then back to the other end 2011-10-23T01:52:29 learnt a new vim shortcut - Ctrl I 2011-10-23T01:52:35 why did i not know this before ==" 2011-10-23T01:52:43 which is totally awesome in maze maps xD 2011-10-23T01:52:50 *** rr5 has joined #aichallenge 2011-10-23T01:53:01 If there weren't so many 1x1 dead ends, I'd suggest a maze solving algorithm 2011-10-23T01:53:07 *** avers_ has joined #aichallenge 2011-10-23T01:53:15 but following walls leads to ants killing eachother without lots of traffic control 2011-10-23T01:53:30 and then, you miss large open areas filled with food 2011-10-23T01:53:32 slows exploration down damn much when they keep running back to the hill continuously :D 2011-10-23T01:53:56 you might be able to do "left wall" with some weighting 2011-10-23T01:54:11 for each water square, weight surrounding squares by 1 2011-10-23T01:54:12 I think I might add an alarm feature 2011-10-23T01:54:23 if square weight > 2 then don't go there 2011-10-23T01:54:30 if an unemployed ant sees an enemy, warn all ants and change their jobs to attack 2011-10-23T01:54:31 xD 2011-10-23T01:54:41 *** he_the_great has joined #aichallenge 2011-10-23T01:54:43 might not be wise :p 2011-10-23T01:55:00 Turn one, I've found no food, but I see an enemy. 2011-10-23T01:55:02 if your strength is say, 10, and you run into an ant, he might have 50 behind him 2011-10-23T01:55:04 Axoren, add an LCD feature >:) 2011-10-23T01:55:07 Drop everything and suicide! 2011-10-23T01:55:10 lol 2011-10-23T01:55:18 LCD? 2011-10-23T01:55:21 Axoren, unemployed ants display statistics by forming numbers and letters! 2011-10-23T01:55:29 xD 2011-10-23T01:55:29 :D 2011-10-23T01:55:37 lol 2011-10-23T01:55:38 troll everyone 2011-10-23T01:55:52 how do you keep track of your ants, axoren 2011-10-23T01:55:52 draw dicks with my ants when food and enemies aren't present 2011-10-23T01:56:07 I don't keep track of them per se. 2011-10-23T01:56:17 I tell them to do something and they ask for another job when they're done 2011-10-23T01:56:25 now i want to see dick formations go into battle with each other 2011-10-23T01:56:30 that implies they have their own state right? 2011-10-23T01:56:35 or stream ants out the end at each other 2011-10-23T01:56:36 *goes to watch replay* -> ants of the other bot write "your bot sucks" 2011-10-23T01:56:59 I have a HashMap where Path is a SinglyLinkedList of Directions 2011-10-23T01:57:37 ah 2011-10-23T01:57:42 When they reach the end of a Path, their Path is null 2011-10-23T01:57:51 When a Path is null, I find something for them to do 2011-10-23T01:58:05 and each ant has their own hashmap? 2011-10-23T01:58:10 wait no that still doesn't work... 2011-10-23T01:58:12 Nope 2011-10-23T01:58:17 The Bot has the hashmap 2011-10-23T01:58:22 The ants are the keys 2011-10-23T01:58:46 yeh that's the part i still don't like, because the keys change making it difficult to keep track ^^ 2011-10-23T01:58:49 when an ant moves, they remove the job mapping and make a new one at their new location 2011-10-23T01:59:40 Think about this: define Path { Path next; Tile/Direction place; } 2011-10-23T01:59:55 nah i get it you put their task in the tile they're going to appear next 2011-10-23T01:59:58 Axoren: what do you do if an enemy ant gets in the way of your path after you've planned it? 2011-10-23T02:00:25 I let the job float 2011-10-23T02:01:08 I could probably do some cleanup at the beginning of the turn to remove floating jobs 2011-10-23T02:01:19 but this leads to an unexpected, nice side effect. 2011-10-23T02:01:29 Another ant will guard the location of a dead ally 2011-10-23T02:01:38 that other ants can come by and pick up the floating job if it comes across it? 2011-10-23T02:01:39 waiting for more enemies, almost 2011-10-23T02:01:48 oh loo 2011-10-23T02:01:49 lol 2011-10-23T02:01:59 aichallenge: McLeopold * r552e598 / worker/worker.py : Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/FN5dcw 2011-10-23T02:02:00 aichallenge: McLeopold * r0d63a91 / (sql/2_generate_matchup.sql sql/opponent.sql): fix to slow matchmaking - http://git.io/mZqGlw 2011-10-23T02:02:06 I could always make jobs inheritable 2011-10-23T02:02:08 I could do lots 2011-10-23T02:02:19 I just want this thing to go into at least one game so I can check food gathering 2011-10-23T02:02:44 Axoren: how do you explore unknown territory? 2011-10-23T02:02:50 Xenophobia 2011-10-23T02:03:07 I missed that part 2011-10-23T02:03:09 it'll happen automatically by having ants run away from eachother when not doing gathering 2011-10-23T02:03:16 oh 2011-10-23T02:03:39 This is my first attempt at any AI 2011-10-23T02:03:43 does that ever push ants into dead ends? 2011-10-23T02:03:45 but I play a lot of video games 2011-10-23T02:03:53 Some times! 2011-10-23T02:04:08 But if they cannot run away from allies 2011-10-23T02:04:08 something to think about then 2011-10-23T02:04:11 they pick a random direction 2011-10-23T02:04:34 aichallenge: McLeopold * r51ae968 / sql/opponent.sql : fix to slow matchmaking - http://git.io/JRW0oQ 2011-10-23T02:04:45 maybe it shouldn't be random, maybe it should be targeted at unknown territory 2011-10-23T02:05:09 That's difficult with my variation of the A* 2011-10-23T02:05:20 I'd have to use an uninformed variant like D* 2011-10-23T02:05:47 let us know when you get a game :D 2011-10-23T02:05:51 wanna watch 2011-10-23T02:06:03 I'll be sure to brag for hours upon hours on end! 2011-10-23T02:06:04 My bot considers unknown area as all land, then replans each turn based on any new information 2011-10-23T02:06:11 Axoren, have you tried the tcp servers? 2011-10-23T02:06:18 I'm very insecure and minor accomplishments are important to me 2011-10-23T02:06:20 TCP? 2011-10-23T02:06:20 Axoren, you can get games there within minutes 2011-10-23T02:06:25 !!! 2011-10-23T02:06:27 Link me! 2011-10-23T02:06:28 ChrisH_ yup that's pretty much how i do it too 2011-10-23T02:06:50 bobbydroptable: I think it's fairly common practice, but I have no way to prove it. 2011-10-23T02:07:02 Alexer, how do! 2011-10-23T02:07:05 Please! 2011-10-23T02:07:23 < ash0d> We have TCP Servers for live games: http://ants.fluxid.pl/howto and http://ash.webfactional.com/howto to connect... 2011-10-23T02:07:54 *** NotCoffeeTable has quit IRC (Quit: Page closed) 2011-10-23T02:08:01 You've made me a happy human being 2011-10-23T02:08:02 oh nice 2011-10-23T02:08:05 McLeopold: hope the fix works and we can get more workers up again. 2011-10-23T02:08:18 i'll definitely run my 2nd bot there 2011-10-23T02:08:47 Axoren, i have my bot running continuously on both :P 2011-10-23T02:09:09 So I need a python runtime, don't I? 2011-10-23T02:09:17 I'm a Win7 user 2011-10-23T02:09:22 Axoren, yep 2011-10-23T02:09:28 Does it come standard on this OS? 2011-10-23T02:09:35 nope 2011-10-23T02:09:46 nope its like java 2011-10-23T02:09:47 I no longer have a tech-ner 2011-10-23T02:09:58 http://www.python.org/ 2011-10-23T02:10:00 I'll be back shortly 2011-10-23T02:10:01 Axoren, you need python to run games locally too 2011-10-23T02:10:02 *** cutoff has quit IRC () 2011-10-23T02:10:09 McLeopold: there seems to be a bug in whether extended games are recorded as different or same ending, but if it's okay with you I'm going to take out the extended recording 2011-10-23T02:10:18 against the sample bots 2011-10-23T02:10:22 *** andy2011 has quit IRC (Quit: Page closed) 2011-10-23T02:10:28 Axoren, or how have you been testing your bot? 2011-10-23T02:10:39 Chalkboard and mentally 2011-10-23T02:10:59 I teach basic java to people on a regular basis, so I have a chalkboard 2011-10-23T02:11:00 Axoren, ouch! 2011-10-23T02:11:05 *** Dominater has joined #aichallenge 2011-10-23T02:11:10 aichallenge: McLeopold * r776edb0 / sql/2_generate_matchup.sql : fix to slow matchmaking - http://git.io/vM3rbQ 2011-10-23T02:11:33 a chalkboard wow 2011-10-23T02:11:35 I remember reading that we can let the bots write to files 2011-10-23T02:11:36 Axoren, i would've hanged myself without local testing :D 2011-10-23T02:11:37 lol 2011-10-23T02:11:43 is there any way to access the files? 2011-10-23T02:11:46 dominater : no you can NOT write to files . 2011-10-23T02:11:47 This isn't as bad as that time I was programming video game graphics in OpenGL at work 2011-10-23T02:12:12 http://aichallenge.org/rules.php 2011-10-23T02:12:13 janzert: okay, matchups much faster now :) 2011-10-23T02:12:16 you can access files you upload 2011-10-23T02:12:19 yay :) 2011-10-23T02:12:22 but they cannot be modified 2011-10-23T02:12:37 any objection to removing extended game recording? 2011-10-23T02:12:55 janzert: plus, i did sort on the wrong column for the first part of the contest, now games should be even 2011-10-23T02:12:55 ... wait? 2011-10-23T02:12:55 Axoren, i had a bug at some point where my pathfinding checked the neighbouring squares for obstructions, then if it found the square on the left to be accessible, it queued the square on the right to be checked :D 2011-10-23T02:13:07 McLeopold: great 2011-10-23T02:13:09 *** yoyoo has quit IRC (Ping timeout: 265 seconds) 2011-10-23T02:13:20 We can use bonus files? 2011-10-23T02:13:34 janzert: you mean, the part where we don't cut off games early and see if it makes a difference? 2011-10-23T02:13:35 Can we somehow detect which map is being used :3? 2011-10-23T02:13:41 Axoren, copypaste error, didn't notice it since 'r' and 'l' are close enough. it took me forever to track it down even with local testing 2011-10-23T02:13:41 McLeopold: right 2011-10-23T02:13:58 axoren: there's no guarantee that the maps will be one of the standard ones? 2011-10-23T02:14:03 I think amstan and I outvoted you on that :) 2011-10-23T02:14:16 an example of the bug is http://aichallenge.org/visualizer.php?game=2695&user=1 it is recorded as different, should be same 2011-10-23T02:14:18 Not so much that I want guarantees that they will be the standard ones 2011-10-23T02:14:29 but we're still recording them atm 2011-10-23T02:14:36 but if I knew which map the bot was on, I could orient more search and destory with more move and destroy :P 2011-10-23T02:14:42 I just want to get rid of that 2011-10-23T02:15:26 just the description? 2011-10-23T02:15:28 I might just make my bots explore just enough to determine which map it is, then load from a file the entire map's layout 2011-10-23T02:15:40 each map's layout being stored in a file 2011-10-23T02:15:42 righ, the game end reason 2011-10-23T02:15:43 :P 2011-10-23T02:15:46 okay 2011-10-23T02:16:14 janzert: did you fire up more workers? 2011-10-23T02:16:31 aichallenge: janzert * r7688e9d / ants/ants.py : Don't record if games are 'extended' - http://git.io/lJ0W9w 2011-10-23T02:16:32 aichallenge: janzert * r1390bc8 / (sql/2_generate_matchup.sql sql/opponent.sql): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/cpQwQg 2011-10-23T02:16:32 that ^^ 2011-10-23T02:16:37 *** djr_ has quit IRC (Ping timeout: 258 seconds) 2011-10-23T02:16:37 not yet I hadn't 2011-10-23T02:16:41 Axoren: you will be disappointed to learn that if they do the same thing as previous competitions there will be all new maps for the finals that you won't know about until after submissions are frozen 2011-10-23T02:16:42 essentially means I can use A* once I know which map it is without fog of war, D* until I do 2011-10-23T02:17:00 *** otzi has quit IRC (Ping timeout: 258 seconds) 2011-10-23T02:17:06 htop looks so much better now :D 2011-10-23T02:17:22 ok, I'll start another woker 2011-10-23T02:17:26 But using a D* until I know what map it is will mean I continue to use D* even if I never find out 2011-10-23T02:17:32 maintaining efficiency 2011-10-23T02:17:44 and worker task results come back nice and fast :) 2011-10-23T02:17:47 all I'd have to do is set a "give up" 2011-10-23T02:17:54 Axoren: my point is that your cache of the "official maps" will be invalidated when they change all the maps for the finals 2011-10-23T02:18:07 ALL of them? 2011-10-23T02:18:10 yep 2011-10-23T02:18:13 :9 2011-10-23T02:18:16 :(* 2011-10-23T02:18:21 they are one step ahead of you 2011-10-23T02:18:22 will there be more competitions in the future? 2011-10-23T02:18:30 hopefully 2011-10-23T02:18:56 Then I'm cool with that 2011-10-23T02:18:57 the idea is to write code that adapts well to its environment, not pre-optimized for a static environment 2011-10-23T02:19:07 janzert: the change looks good 2011-10-23T02:19:47 McLeopold: the games are clicking by much faster even with one worker, nice job. 2011-10-23T02:19:57 ChrisH_: keep your eye on the game rates for me 2011-10-23T02:20:04 new worker request in 2011-10-23T02:20:43 McLeopold: no problem, already up to 1.8 per min over the last 5, before it was below 0.5 2011-10-23T02:20:55 4.7 for me 2011-10-23T02:21:01 speedy 2011-10-23T02:21:57 *** Nakyuant has quit IRC (Quit: Page closed) 2011-10-23T02:22:01 *** k333es has joined #aichallenge 2011-10-23T02:22:17 *** aciobanu has quit IRC (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org) 2011-10-23T02:22:36 bobbydroptable: that's players per minute, i was using games per minute from this page: http://aichallenge.org/server_stats.php 2011-10-23T02:22:51 ah 2011-10-23T02:23:32 where does the installer put python and how do I change the PATH in Windows? 2011-10-23T02:23:32 it's going back down, though :( 2011-10-23T02:23:39 janzert: McLeopold the page navigation links on the standings page are all wacky 2011-10-23T02:23:47 forum/forum? 2011-10-23T02:23:49 apparenly, the installer did everything except that 2011-10-23T02:24:05 I've seen this several times before, and it always seems to go away, but I have no idea what causes it 2011-10-23T02:24:14 me neither :( 2011-10-23T02:24:16 ChrisH_: hmm, they appear fine for me 2011-10-23T02:24:21 I saw it once yesterday 2011-10-23T02:24:26 Axoren, i think it was C:\PythonXY 2011-10-23T02:24:29 click page 16 then prev 2011-10-23T02:24:34 not sure though 2011-10-23T02:24:50 ooh, page 1 has no css for some reason 2011-10-23T02:24:51 janzert: it's gone now 2011-10-23T02:25:02 janzert: check your links now 2011-10-23T02:25:32 or I see the url is all messed up 2011-10-23T02:25:37 http://aichallenge.org/rankings.php/forum/forum/forum/forum/forum/forum/forum/country_profile.php?page=1 2011-10-23T02:25:40 I think the lack of formatting is a symptom, though, cause that's the result of click the bad links 2011-10-23T02:26:00 yeah, that's the problem I was referring to 2011-10-23T02:26:04 I think amstan had a redirect for /forum 2011-10-23T02:26:14 I would guess it's related 2011-10-23T02:26:42 $targetpage = $_SERVER['PHP_SELF']; ? 2011-10-23T02:27:13 C:\Python27>python.exe tcpclient.py ants.fluxid.pl 2081 "java -jar M:\Ants\MyBot .jar" axoren XXXXXX 1 2011-10-23T02:27:37 Getting errors in jar execution, not being able to find the main class 2011-10-23T02:27:47 also, that space is from copying 2011-10-23T02:27:56 *** kronos has joined #aichallenge 2011-10-23T02:28:49 *** korfuri has left #aichallenge 2011-10-23T02:28:51 janzert: is the redirect in the apache settings? 2011-10-23T02:28:56 Axoren, does it work if you do just "C:\Python27>java -jar M:\Ants\MyBot.jar"? 2011-10-23T02:29:16 McLeopold: I thought so, but now that I look at them I don't see that he has anything modified there 2011-10-23T02:29:27 Axoren, as in, it should start without errors but not give any output (since it doesn't get input..) 2011-10-23T02:29:45 janzert: can you reproduce it? I think it's a special combination of clicks 2011-10-23T02:30:12 gotit 2011-10-23T02:30:15 or memcache is getting messed up somehow? 2011-10-23T02:30:38 memcache? 2011-10-23T02:30:43 McLeopold: I have not found it to be reliably reproducable. 2011-10-23T02:30:45 how would that do it? 2011-10-23T02:30:58 I mean the information getting put into it 2011-10-23T02:31:03 *** webchat has joined #aichallenge 2011-10-23T02:31:11 it would explain why it seems to only last for a short time then go away 2011-10-23T02:31:25 *** Nakyuant has joined #aichallenge 2011-10-23T02:31:29 *** webchat is now known as Guest86160 2011-10-23T02:31:32 but it does seem vaguely like it happens around the time of code changes, but that could be selective memory 2011-10-23T02:31:35 on my part 2011-10-23T02:31:45 if you see it, do a view source and pastebin for me 2011-10-23T02:32:12 ok 2011-10-23T02:32:50 of course I don't have a pastebin account. ;) 2011-10-23T02:32:58 you don't need one 2011-10-23T02:33:21 does pastebin have accounts? 2011-10-23T02:33:21 ok, just leave you the url here? 2011-10-23T02:33:31 it has a sign in button 2011-10-23T02:33:41 yes 2011-10-23T02:33:56 heh, never noticed :) 2011-10-23T02:34:14 2nd worker is up 2011-10-23T02:34:22 *** heeltoehero has quit IRC (Quit: Lost terminal) 2011-10-23T02:35:42 janzert: McLeopold question: with the backlog of games there are lots of accounts with 0 games played on submission #1, from what you have described I would expect all games to have at least one of these players in them because they should all be at the front of the queue, but that is not true for the recent games 2011-10-23T02:35:57 janzert: McLeopold so what's the explanation? 2011-10-23T02:36:25 something else is broken? :/ 2011-10-23T02:36:29 :( 2011-10-23T02:37:04 *** kronos has quit IRC (Disconnected by services) 2011-10-23T02:37:16 *** kronos has joined #aichallenge 2011-10-23T02:39:05 McLeopold: still cranking through games at a much improved speed though :) 2011-10-23T02:39:43 *** murilo has quit IRC (Quit: fui...) 2011-10-23T02:40:17 the second worker doesn't seem to be doing much though 2011-10-23T02:41:40 ChrisH_: according to my query, all of the last five games include mostly players with 0 or 1 games 2011-10-23T02:41:46 ChrisH_: new workers are slow because they don't have any of the bots downloaded and compiled yet 2011-10-23T02:42:08 janzert: ok, i'll keep that in mind 2011-10-23T02:42:12 *** diego_diego has left #aichallenge 2011-10-23T02:42:26 *** Puj has joined #aichallenge 2011-10-23T02:42:27 McLeopold: players or submissions? 2011-10-23T02:42:44 http://aichallenge.org/visualizer.php?game=2755 seems to have only players with multiple games 2011-10-23T02:43:31 4 players, 3 with 0 games 2011-10-23T02:43:36 by user 2011-10-23T02:43:42 McLeopold: every player in this game had at least 6 games before it started http://aichallenge.org/visualizer.php?game=2749 2011-10-23T02:43:59 McLeopold: all their profile pages show many more games 2011-10-23T02:44:57 *** djr_ has joined #aichallenge 2011-10-23T02:45:54 minimum seems to be 8 games (2 players with 8, 1 with 13, 1 with 14) 2011-10-23T02:47:08 but one of the players had a submission with 0 games, so I think that is a clue 2011-10-23T02:47:57 okay, I see the issue 2011-10-23T02:48:12 yay! 2011-10-23T02:48:57 oh, did you look it the games were in the last 24? 2011-10-23T02:49:17 that's the problem 2011-10-23T02:49:22 but it's not a problem 2011-10-23T02:49:23 :) 2011-10-23T02:49:24 *** djr_ has quit IRC (Client Quit) 2011-10-23T02:49:26 they are 2011-10-23T02:49:29 no, that's hard to do just using the web pages 2011-10-23T02:49:55 give me a user_id that you think has lots of games 2011-10-23T02:50:27 I have a feeling you are right McLeopold , but I'll keep looking around. 2011-10-23T02:50:44 we've had so few games, that most everyone is considered as having "no games" as far as keeping it even goes 2011-10-23T02:50:50 ahh, no. not all of them had games within 24 hours 2011-10-23T02:50:52 in the last 24 hours 2011-10-23T02:51:09 but seeds shouldn't be limited to 24 hour games should they? 2011-10-23T02:51:17 that means their lower sigma, which is 3rd in line, is taking over 2011-10-23T02:51:19 although it shouldn't make a difference normally 2011-10-23T02:51:28 janzert: no, seeds are not 2011-10-23T02:51:35 seeds are least recently played 2011-10-23T02:51:40 so there is still a problem 2011-10-23T02:51:49 that's why you should have at least one in there that looks good 2011-10-23T02:51:51 never played should be less recent than anything 2011-10-23T02:51:55 yes 2011-10-23T02:52:05 which game should I look at? 2011-10-23T02:52:13 http://aichallenge.org/visualizer.php?game=2755 2011-10-23T02:52:20 is the one I'm looking at 2011-10-23T02:52:40 seed_id was 1793, with 3 games in last 24 2011-10-23T02:52:50 :( 2011-10-23T02:53:08 and that's the player with the new submission 2011-10-23T02:53:32 yep 2011-10-23T02:53:34 hmmm 2011-10-23T02:53:41 sounds fishy. ;) 2011-10-23T02:54:09 and his previous game was just minutes for that game too 2011-10-23T02:54:13 *** he_the_great has quit IRC (Ping timeout: 240 seconds) 2011-10-23T02:54:29 yep, problem is in seed query 2011-10-23T02:54:35 I know the issue...fixing 2011-10-23T02:54:46 limits the submission to latest and then orders by submission max_game_id 2011-10-23T02:54:48 ok 2011-10-23T02:55:06 yeah, max_game_id used to be a query across all the games 2011-10-23T02:55:24 I switched it for speed, but it had an unintended consequence :( 2011-10-23T02:55:44 if you can find the latest submission with a game it would work 2011-10-23T02:56:02 it should be any game, I just need to make sure it is fast 2011-10-23T02:56:31 while were here debugging, I found the line of code that produces the weird (>") rank for new submissions on the first page of ranks, but I'm not enough of a .php expert, nor do I know your web architecture well enough to fix it. 2011-10-23T02:56:51 ChrisH_: it's a query issue 2011-10-23T02:56:55 those wierd ranks are called ghosts 2011-10-23T02:57:06 what causes them? disqualification? 2011-10-23T02:57:09 they just shouldn't be showing up for new players 2011-10-23T02:57:14 How do I fix: "Unable to access jarfile MyBot" when using the TCP server? 2011-10-23T02:57:18 yeah, and when they are on there you can't use the sort columns either 2011-10-23T02:57:34 it looks like bad HTML is produced 2011-10-23T02:57:39 *** J0KRESTER has quit IRC (Quit: Page closed) 2011-10-23T02:57:43 by bad, I mean not well formed 2011-10-23T02:57:49 *** Puj has quit IRC (Ping timeout: 258 seconds) 2011-10-23T02:57:51 ChrisH_: they were a feature I removed 2011-10-23T02:58:02 hopefully they will come back correctly 2011-10-23T02:59:35 well, line 253 of website / ranking.php looks suspicious to me 2011-10-23T03:00:20 http://pastebin.com/ZH4ZdxQG 2011-10-23T03:00:46 line 253 is a blank line? 2011-10-23T03:00:48 the tag isn't closed 2011-10-23T03:01:00 ahh, I see it 251 2011-10-23T03:01:05 yes, I did write that, it is kirby the ghost, meant to represent submissions that have been replaced 2011-10-23T03:01:12 i'm going from the line #s on github 2011-10-23T03:01:37 I should replace with > 2011-10-23T03:01:41 there is no 2011-10-23T03:01:42 and close the td tag :( 2011-10-23T03:01:52 git it 2011-10-23T03:01:55 got it even :) 2011-10-23T03:01:56 got it 2011-10-23T03:02:05 ok I'll let you have it 2011-10-23T03:02:15 no you, I'm fixing matchups 2011-10-23T03:02:47 aichallenge: janzert * r483ff71 / website/ranking.php : Fix bad ghost markup - http://git.io/6oSnLg 2011-10-23T03:02:54 so maybe I'm lame, but who is kirby the ghost? 2011-10-23T03:03:22 pro tip: don't type sql into the linux command line :/ 2011-10-23T03:04:00 *** kronos has quit IRC (Ping timeout: 240 seconds) 2011-10-23T03:04:01 *** burny has joined #aichallenge 2011-10-23T03:04:14 http://www.google.com/search?q=kirby+ghost 2011-10-23T03:05:40 gah, fixed the html the wrong way 2011-10-23T03:05:58 haha 2011-10-23T03:06:55 aichallenge: janzert * re697a6a / website/ranking.php : Really fix ghost markup - http://git.io/51xG0g 2011-10-23T03:06:59 try that again 2011-10-23T03:07:54 I'm not sure how kirby ends up as (>")> though 2011-10-23T03:07:59 so kirby is from some video game I've never heard of and the ghost is one of his special powers, got it 2011-10-23T03:08:19 well I guess I kind of see it :} 2011-10-23T03:08:29 but do you want the ghosted entries to show up at the top of the standings? 2011-10-23T03:08:35 I thought it was a well known emoticon 2011-10-23T03:08:44 ChrisH_: no, that's a bug 2011-10-23T03:08:54 ChrisH_: they are suppose to show up where the users previous submission was ranked 2011-10-23T03:09:12 ahh, that explains the text on the markup 2011-10-23T03:09:13 the bug is when the user doesn't have a previous rank 2011-10-23T03:09:15 *** rr5 has quit IRC (Remote host closed the connection) 2011-10-23T03:09:41 McLeopold: I had never run into it before 2011-10-23T03:09:42 *** Dominater has quit IRC (Ping timeout: 265 seconds) 2011-10-23T03:10:15 *** kronos has joined #aichallenge 2011-10-23T03:13:52 aichallenge: McLeopold * r571c617 / (sql/2_generate_matchup.sql sql/opponent.sql): fix wrong seed player - http://git.io/WtOFEg 2011-10-23T03:13:52 aichallenge: McLeopold * r9cb8bf9 / (ants/ants.py website/ranking.php): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/dcX-9g 2011-10-23T03:15:04 any luck on getting a weird pagination yet? 2011-10-23T03:15:18 not recently 2011-10-23T03:15:22 I haven't been able to 2011-10-23T03:17:48 McLeopold: you keep changing the indentation of sql files, it's making it hard to tell what actually changed :} 2011-10-23T03:17:58 *** iris1 has quit IRC (Read error: Connection reset by peer) 2011-10-23T03:18:04 aichallenge: McLeopold * r02b1c16 / website/sql.php : fix extra kirbies - http://git.io/0uO3-Q 2011-10-23T03:18:10 sry, I blame eclipse 2011-10-23T03:18:37 it was tabs to spaces 2011-10-23T03:19:15 games are slowing down as we get a bunch of the top bots playing for the first time in many hours, but there are some good games to watch. :) 2011-10-23T03:20:13 janzert: did we have enough food here? http://aichallenge.org/visualizer.php?game=2817 2011-10-23T03:20:24 *** stano has quit IRC (Ping timeout: 260 seconds) 2011-10-23T03:21:37 seems rather slow, but it was still an interesting game 2011-10-23T03:21:52 so it doesn't seem to bad from that case at least 2011-10-23T03:21:58 FlagCapper takes down MomoBot and Sebaplus, http://aichallenge.org/visualizer.php?game=2812&user=31 2011-10-23T03:22:24 sir_macelon could have pushed through and ended the game pretty easily 2011-10-23T03:22:41 anytime form around move 600 on 2011-10-23T03:22:54 mleise: you around? 2011-10-23T03:23:14 latest seed changes live? 2011-10-23T03:23:25 should be 2011-10-23T03:23:35 *** iris1 has joined #aichallenge 2011-10-23T03:23:43 ok 2011-10-23T03:23:55 most everyone has games, that has a valid bot 2011-10-23T03:24:01 low food games are the most challenging strategically, IMHO 2011-10-23T03:24:36 oh, wow it ate through the 200+ players quick 2011-10-23T03:25:32 Well, I've given up on bug fixing for tonight 2011-10-23T03:25:44 but apparently, the batch file provided whe nthe java starter file 2011-10-23T03:25:45 hmm, both servers are running real slow now 2011-10-23T03:25:50 made an empty jar 2011-10-23T03:25:58 and that's what was uploaded to the site 2011-10-23T03:26:09 some impressive micro in that game w/ flagcapper 2011-10-23T03:26:27 guess i'll need to implement battle stuff 2011-10-23T03:26:33 a1k0n: both playing a game with many players 2011-10-23T03:26:44 *** AlvinMinring has joined #aichallenge 2011-10-23T03:27:23 janzert: it would be nice if the maps page was in table format, then I could sort 2011-10-23T03:28:05 also, do you mind if I change the game reasons and remove the extended part? 2011-10-23T03:28:36 doesn't seem like all the information would fit in a table particularly well 2011-10-23T03:28:47 McLeopold: I already did 2011-10-23T03:28:53 or you mean in the database? 2011-10-23T03:29:02 *** xar0l has quit IRC (Ping timeout: 265 seconds) 2011-10-23T03:29:12 a1k0n: games with top bots take longer too because they spend more time computing each turn compared to the starter bots 2011-10-23T03:29:16 I don't mind changing 2011-10-23T03:29:19 21 columns? 2011-10-23T03:29:23 McLeopold : back to back games 2817 and 2818. almost same opponents. not sure whether it can be a clue to a some issue or perfectly normal :-) 2011-10-23T03:29:27 ChrisH_: right 2011-10-23T03:29:44 *** Axoren has quit IRC (Quit: Page closed) 2011-10-23T03:31:47 McLeopold: looks like new submissions aren't getting games right away anymore, so that fix seems to work 2011-10-23T03:32:31 *** boegel has joined #aichallenge 2011-10-23T03:32:39 hiya y'all 2011-10-23T03:32:52 seems like a 21 column table will be a wall of text, quite hard to read. but sorting would be nice too 2011-10-23T03:32:58 ashoka: that does seem off a little 2011-10-23T03:33:15 can someone update the Haskell starter package on the website? there's a version with hills support now (see https://github.com/qacek/ants-haskell) 2011-10-23T03:33:49 *** klasbj has joined #aichallenge 2011-10-23T03:34:03 boegel: is there a pull request to get that back into the main repo? 2011-10-23T03:34:20 it doesn't look forked 2011-10-23T03:34:53 arr darn i lost to ChrisH 2011-10-23T03:35:12 hehe 2011-10-23T03:35:38 *** Zacru has joined #aichallenge 2011-10-23T03:36:00 hehehe no fuss 2011-10-23T03:36:00 bobbydroptable: looks like that was your first time not fiinishing first too 2011-10-23T03:36:06 nah 2nd 2011-10-23T03:36:18 its done surprisingly well 2011-10-23T03:36:25 considering it doesn't even attack 2011-10-23T03:36:26 lol 2011-10-23T03:36:41 McLeopold : just 1 of the 5 opponents had changed. Map is the same too, 1-2 minutes apart. hmmm 2011-10-23T03:37:08 *** crusher has joined #aichallenge 2011-10-23T03:38:18 boegel: read this -> https://github.com/aichallenge/aichallenge/wiki/Ants-Starter-Pack-Guide, then, could you implement unoccupied and anything else that might be missing? 2011-10-23T03:38:40 bobbydroptable: yeah, my bot doesn't know how to micro, it just tries to bull rush your hills and hope to have overwhelming numbers 2011-10-23T03:38:56 yeh 2011-10-23T03:38:57 *** Dominater has joined #aichallenge 2011-10-23T03:39:06 well i have more personal goals i want to accomplish with the bot 2011-10-23T03:39:08 not just win 2011-10-23T03:39:09 lol 2011-10-23T03:39:24 sigh i still gotta go ml/ai class hwk 2011-10-23T03:39:26 I still think that for a game like this, having teams would be more interesting 2011-10-23T03:39:28 i'm still thinking through how to implement combat at the micro level 2011-10-23T03:39:49 *** vladimirfol has joined #aichallenge 2011-10-23T03:39:57 well the immediate thing i can think of 2011-10-23T03:40:09 is for ants to stay close to your own ants 2011-10-23T03:40:24 that's probably the easiest way 2011-10-23T03:40:39 yeah, but orientation matters too 2011-10-23T03:40:49 yeh 2011-10-23T03:40:50 is the only way to win this: to kill the most hills? 2011-10-23T03:40:56 burny: yes 2011-10-23T03:40:57 yep 2011-10-23T03:41:04 i highly doubt we'll see any significant formation implementations 2011-10-23T03:41:11 and you kill a hill, just by steping on it? 2011-10-23T03:41:16 burny yup 2011-10-23T03:41:32 burny: do the specs need to be more clear? 2011-10-23T03:41:44 Hi, I was hoping someone could help me to get my code to compile with C++11? (STL hash sets are nice...) According to the README in the starter pack the server should try to use the Makefile I submit in my zip file, where I have included the -std=c++0x flag, but it doesn't compile, because the -std=c++0x flag is not present. Is it not possible to use C++11 for this challenge, or is there some was around it? 2011-10-23T03:42:31 they're clear enoug hI suppose, it just seems odd, that creating an army, and killing ants, doesn't give points 2011-10-23T03:42:40 janzert: looks like we lost a worker 2011-10-23T03:42:42 klasbj: the server does not use your makefile 2011-10-23T03:42:45 yeh maybe there should have been a bonus for having the most ants at the end of the game 2011-10-23T03:42:48 how is that odd? 2011-10-23T03:42:59 McLeopold: workers are getting blank tasks 2011-10-23T03:43:22 'seems' odd 2011-10-23T03:43:25 now they are both down 2011-10-23T03:43:45 McLeopold: unoccupied is not there yet for the Haskell starter bot? 2011-10-23T03:43:56 ChrisH_: they're just unable to get tasks from the main server 2011-10-23T03:44:22 oh, so if they're rate drops too low they disappear from the stats page? 2011-10-23T03:44:41 yep, if the rate drops to 0 they're gone 2011-10-23T03:44:48 ok 2011-10-23T03:44:52 janzert: Ok, I figured as much, a bit annoying when the readme in the starter pack says it should... Do you know if there is any way to tell the server to compule it with certain flags? 2011-10-23T03:45:15 *** Dominater has quit IRC (Quit: Page closed) 2011-10-23T03:45:23 klasbj: no that is fixed in the server code 2011-10-23T03:45:43 fixed as in static, unchanging 2011-10-23T03:45:53 right 2011-10-23T03:46:21 janzert: can you pause the workers? 2011-10-23T03:46:26 *** crusher has quit IRC (Ping timeout: 265 seconds) 2011-10-23T03:46:27 sure 2011-10-23T03:46:30 janzert: no, there's no pull request afaik 2011-10-23T03:47:10 *** SofiaWhite has joined #aichallenge 2011-10-23T03:47:18 janzert: but, the Haskell starter package in the main repo is just a synlink to the ant-haskells github repo? 2011-10-23T03:47:24 janzert: not sure how that works... 2011-10-23T03:47:52 ahh, the main repo just needs the submodule to be updated then 2011-10-23T03:48:12 I can never remember how to do that though :/ 2011-10-23T03:48:17 janzert: and how should I trigger that? 2011-10-23T03:48:20 janzert: ah :) 2011-10-23T03:48:27 ask amstan is the best way ;) 2011-10-23T03:48:39 janzert: also, seems like some more work is needed for the starter package to be complete, like McLeopold pointed out 2011-10-23T03:49:56 seeing I know pretty much 0 haskell I don't have a chance of helping with that 2011-10-23T03:50:41 Hi guys 2011-10-23T03:50:53 hello 2011-10-23T03:51:12 Are you aware of any trick to better debug the bot? 2011-10-23T03:51:22 output debugging messages to error output 2011-10-23T03:51:26 which language? 2011-10-23T03:51:49 I'm using C# and it seems there's no way to trace/debug live? 2011-10-23T03:52:26 http://msdn.microsoft.com/en-us/library/system.console.error.aspx 2011-10-23T03:52:32 Well, I don't even see the bot console window 2011-10-23T03:52:39 oh 2011-10-23T03:52:49 i don't have the C# starter pack so i don't know how it is done 2011-10-23T03:52:56 so you don't run it on the command line? 2011-10-23T03:53:10 *** Zacru has quit IRC (Quit: Page closed) 2011-10-23T03:53:11 do you run it in visual studio? 2011-10-23T03:53:43 hmm, where does the standard error output output when you don't see the app window and can't attach your debugger? 2011-10-23T03:54:10 SofiaWhite: the engine will capture the bots stderr and output it to the same console as the engine is being run in 2011-10-23T03:54:11 I run it both from Windows Explorer and command prompt 2011-10-23T03:54:27 it runs as a console app I believe? 2011-10-23T03:54:33 if you run it from command prompt 2011-10-23T03:54:51 I see janzert. I guess that's what bobby is also saying. I'll try that and come back. Thanks 2011-10-23T03:55:21 yep, it does. I'll try to output to stderr 2011-10-23T03:55:28 brb 2011-10-23T03:57:58 *** Pawel__ has joined #aichallenge 2011-10-23T03:58:14 rugby worldcup about to start 2011-10-23T03:58:23 in case anyone is too engrossed in bot 2011-10-23T03:58:51 hello - tutorial says how to test our bouts - it says to do this file tutorial.sh by pasting this loooong comand:user@localhost:~gedittutorialshAftereditingthefilemakeitexecutabletextuserlocalhost oh ... 2011-10-23T04:00:12 yeah ok - thanks 2011-10-23T04:00:37 question. The current game rate is about 0.3 players per minute. is it ok?can i help with my dualcore to speed it up??? 2011-10-23T04:00:46 dcheva lol no 2011-10-23T04:00:54 all the games are run on the server 2011-10-23T04:01:00 *** Pawel__ has quit IRC (Client Quit) 2011-10-23T04:01:21 nothing is done on your computer after you have uploaded the files 2011-10-23T04:01:26 google have not normal server... understand. 2011-10-23T04:01:46 What's the difference between rank and skill? 2011-10-23T04:01:48 *** Nakyuant has quit IRC (Quit: Page closed) 2011-10-23T04:01:57 skill level determines rank, no? 2011-10-23T04:02:10 rank is just the numerical index of the order of skill levles 2011-10-23T04:02:13 I have no idea. Sometimes one goes up while the other goes down 2011-10-23T04:02:36 How is skill determined? 2011-10-23T04:02:49 apparently, http://research.microsoft.com/en-us/projects/trueskill/ 2011-10-23T04:03:02 don't forget your rank can fall if someone gets a higher skill level 2011-10-23T04:03:08 even if your skill went up 2011-10-23T04:04:27 ok bb 2011-10-23T04:04:29 *** nullkuhl has quit IRC (Ping timeout: 252 seconds) 2011-10-23T04:04:30 *** dcheva has quit IRC (Quit: Page closed) 2011-10-23T04:05:41 *** danielharan has joined #aichallenge 2011-10-23T04:05:55 aichallenge: McLeopold * rf5b4023 / (sql/2_generate_matchup.sql sql/opponent.sql): fix average game count - http://git.io/W6NkHg 2011-10-23T04:06:07 hello! i just want to ask. How you solve fast calculation ants ways problem? I think i can create graph of map areas(each 5 cells i think) and connections of them. 2011-10-23T04:06:20 i am trying that too 2011-10-23T04:06:24 but iti s proving difficult 2011-10-23T04:06:35 because paths may not fit into cell 2011-10-23T04:06:38 for example 2011-10-23T04:06:53 if you have a 5 cell area, separated by water 2011-10-23T04:06:58 janzert: back in business 2011-10-23T04:07:04 great 2011-10-23T04:07:07 you can have areas in that cell that cannot find path to the center of the area 2011-10-23T04:07:32 for calculating any form of paths, suggest looking at A* for rudimentary algorithm 2011-10-23T04:09:18 *** Pawel__ has joined #aichallenge 2011-10-23T04:09:24 i think about it, in every area will be only cell that visible for ant (if he stau in center of area) and that possible to get (in 5 steps) 2011-10-23T04:11:40 each area will know about neighbors, if ant stay to cell that not belong to any area, than i will create new area with center in this cell, calculate cell belongs to the area and insert the area to the graph 2011-10-23T04:11:43 sorry guys - but i am not very good with linux :-). In this tutorial on how to test our bots - it says to create some script by pasting a command line. I guess it opens up an editor of some sort were i supposed to write the lines that are given - simple. But how on earth can I save it and go back to the shell ?? 2011-10-23T04:11:44 *** dungeonduke has quit IRC (Remote host closed the connection) 2011-10-23T04:11:49 game servers still stale? 2011-10-23T04:11:50 i am sorry abou my English BTW 2011-10-23T04:11:55 aichallenge: Nick Ham * rc311f10 / website/specification.php : Fixed adding a section to the specs page on food harvesting - http://git.io/d0hw1g 2011-10-23T04:12:19 *** Puj has joined #aichallenge 2011-10-23T04:12:27 cirno_the_greate: they are spooling back up 2011-10-23T04:12:52 antimatroid: that looks like it just duplicated a section? 2011-10-23T04:13:10 janzert: i added it to the contents but not the actual content page bit 2011-10-23T04:13:21 http://aichallenge.org/specification.php see there 2011-10-23T04:13:25 cutoff rules repeats twice 2011-10-23T04:13:27 but it says cuttof rules 2011-10-23T04:14:01 oh, i missed a bit, sorry :P 2011-10-23T04:14:14 :) 2011-10-23T04:14:47 aichallenge: Nick Ham * r1303b93 / website/specification.php : Fixed adding a section to the specs page on food harvesting again - http://git.io/iIY-Vg 2011-10-23T04:15:03 sorry guys - but i am not very good with linux :-). In this tutorial on how to test our bots - it says to create some script by pasting a command line. I guess it opens up an editor of some sort were i supposed to write the lines that are given - simple. But how on earth can I save it and go back to the shell ?? 2011-10-23T04:15:32 Pawel__: do you have gedit open? 2011-10-23T04:15:43 Pawel__: open up a shell window, cd to your ants directory and run the play game script :P 2011-10-23T04:15:53 it isnt gedit? is within the black shell screan 2011-10-23T04:15:58 woah, the linux script is bad on the tutorial 2011-10-23T04:15:59 McLeopold: ready for another worker to come online? 2011-10-23T04:16:04 janzert: sure 2011-10-23T04:16:22 Pawel__: the tutorial page is really messed up 2011-10-23T04:16:26 ok, third worker coming up. then I gotta get to bed 2011-10-23T04:16:37 oh ok - so what do u suggest? 2011-10-23T04:16:53 McLeopold: match scheduling looks like it is behaving more like intended now. :) 2011-10-23T04:16:59 https://github.com/aichallenge/aichallenge/wiki/Ants-Tutorial 2011-10-23T04:17:07 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T04:17:07 ChrisH_: yeah! 2011-10-23T04:17:16 Pawel__: read that link 2011-10-23T04:17:30 Thanks - i'll go through it :-) 2011-10-23T04:17:39 *** Pawel__ has quit IRC (Quit: Page closed) 2011-10-23T04:17:39 any regex experts? 2011-10-23T04:17:49 *** bitbrit has joined #aichallenge 2011-10-23T04:18:32 McLeopold: when the workers came back on line they quickly played through the first time submissions, and now it looks like each game has a player that hasn't had a game in a really long time. 2011-10-23T04:18:53 McLeopold, you might have better luck in #regex 2011-10-23T04:19:09 McLeopold: could you take a look into this traceback http://pastebin.com/x8nVtw0e 2011-10-23T04:19:11 players with games and recent submissions are not jumping the queue 2011-10-23T04:19:31 that's from a failed matchup in the matchup table earlier 2011-10-23T04:19:35 McLeopold: might be able to help a bit with the regex 2011-10-23T04:19:57 janzert: that's weird, a -1 coord? 2011-10-23T04:20:03 Heys guys, are the unofficial(?) TCP servers (Fluxid, ash0d) listed somewhere? 2011-10-23T04:20:12 @tcp 2011-10-23T04:20:13 McLeopold: I do not know about 'tcp', but I do know about these similar topics: 'topic' 2011-10-23T04:20:23 Puj: no, they should be 2011-10-23T04:20:25 the thing is how did it get past the order validation 2011-10-23T04:20:41 janzert: do we have any logs of the game? 2011-10-23T04:20:44 or where else did it come from? 2011-10-23T04:20:47 like bot input or output? 2011-10-23T04:20:58 nope, just that 2011-10-23T04:20:58 Aye. Guess that's not highest prio. 2011-10-23T04:20:59 gah my bot is falling like a fly in the rankings 2011-10-23T04:21:01 lol 2011-10-23T04:21:20 and it is the first time I've seen that traceback 2011-10-23T04:21:36 do we know which bots and such? 2011-10-23T04:21:53 it's 2624 in the matchup table 2011-10-23T04:21:59 err, 3624 2011-10-23T04:23:16 janzert: my line numbers don't match :/ 2011-10-23T04:23:30 also it might be nice if you could delete all the matchups with "abort matchup: could not find opponent" 2011-10-23T04:23:44 oh it happened before taking out the extended game stuff 2011-10-23T04:24:02 let me find the new location of that 2011-10-23T04:24:12 *** avers_ has joined #aichallenge 2011-10-23T04:24:34 it would be line 626 now 2011-10-23T04:24:49 why the difference? 2011-10-23T04:24:55 Puj: you can list them in topic at forum ;) 2011-10-23T04:25:03 because I removed the extended game recording 2011-10-23T04:25:07 ok 2011-10-23T04:25:25 https://github.com/aichallenge/aichallenge/commit/7688e9d334a9c0762215f8d8cd70035ea67e9cfc 2011-10-23T04:25:28 that commit 2011-10-23T04:25:54 *** b4Ant has quit IRC (Remote host closed the connection) 2011-10-23T04:27:04 cirno_the_greate, is the topic created or are you suggesting I create it? Unfortunately I don't have the list of servers. 2011-10-23T04:27:18 @help topic 2011-10-23T04:27:18 McLeopold: (topic []) -- Returns the topic for . is only necessary if the message isn't sent in the channel itself. 2011-10-23T04:27:24 @help factoid 2011-10-23T04:27:24 McLeopold: Error: There is no command "factoid". 2011-10-23T04:27:38 @help learn 2011-10-23T04:27:38 McLeopold: (learn [] as ) -- Associates with . is only necessary if the message isn't sent on the channel itself. The word 'as' is necessary to separate the key from the value. It can be changed to another word via the learnSeparator registry value. 2011-10-23T04:27:38 janzert: McLeopold ok, another weird web page behavior, the rankings page is now showing only users with org "Other" 2011-10-23T04:28:07 *** Palmik has quit IRC (Remote host closed the connection) 2011-10-23T04:28:12 janzert: McLeopold I've seen this a few times before and don't know if it is related to the forum/forum/forum link stuff or not 2011-10-23T04:28:23 *** Palmik has joined #aichallenge 2011-10-23T04:28:30 doesn't reproduce for me atm :( 2011-10-23T04:28:52 I can't select only other :/ 2011-10-23T04:28:55 must be the zero 2011-10-23T04:29:05 that I am seeing 2011-10-23T04:29:06 no, it dissappeared quick, I only saw it once 2011-10-23T04:29:39 McLeopold: someone also mentioned the other day that some organization pages are broken, such as 2011-10-23T04:29:40 but when it happens it is always filtered to "other" 2011-10-23T04:29:40 http://aichallenge.org/organization_profile.php?org=921 2011-10-23T04:30:00 janzert: yeah, any org with one user shows a page saying invalid org 2011-10-23T04:30:17 http://aichallenge.org/organization_profile.php?org=993 2011-10-23T04:30:21 I tried looking into it but couldn't figure out was going wrong, before I got distracted by some other problem :/ 2011-10-23T04:30:46 *** Pawel__ has joined #aichallenge 2011-10-23T04:31:00 ChrisH_: this works though http://aichallenge.org/organization_profile.php?org=356 2011-10-23T04:31:17 interesting 2011-10-23T04:31:18 Guys do you have any info for volunteers interested in helping out? 2011-10-23T04:31:30 that error should only happen before it is at the stage of even being able to tell how many users an org has 2011-10-23T04:31:44 i.e. it hasn't done the query on the org users yet 2011-10-23T04:32:49 Here is another: http://aichallenge.org/organization_profile.php?org=921 2011-10-23T04:32:59 oh, same one janzert posted 2011-10-23T04:33:04 :) 2011-10-23T04:33:05 *** Belerafon_L has joined #aichallenge 2011-10-23T04:33:14 which is the one I sent you ;) 2011-10-23T04:33:42 heh, now it works 2011-10-23T04:33:48 *** Macuyiko has joined #aichallenge 2011-10-23T04:34:20 ha, here is one with no entries, but a good page: http://aichallenge.org/organization_profile.php?org=994 2011-10-23T04:34:38 924 as well 2011-10-23T04:34:49 Hello guys - really sorry for trouble - that's me again: I went through tutorial and i've done it - it works (The test) - but it throughs exceptions and cannot find my class "MyBot" ("Could not find the main class: MyBot. Program will exit.") - I checked it was there - so why it could not find it? should I add it to my path or something? 2011-10-23T04:35:43 that looks like a language runtime error? 2011-10-23T04:36:15 i've only submitted the starter program and i'm 850 out of ~1600. is that just chance, or have hundreds of people submitted bots worse than the starter bot? 2011-10-23T04:36:17 yes - loads of java exceptions - but all are ClassNotDefined error type 2011-10-23T04:36:23 Hi. I'm a new. I have compile C++ version of ants, but can't run it. That parameters line I need to set on terminal? (loadtime, turntime and etc.) 2011-10-23T04:36:44 Belerafon_L, that's passed in on stdin 2011-10-23T04:36:45 Set your classpath to the dir where your Java files are and run java full_path_to_MyBot 2011-10-23T04:37:21 ok thanks - i will do it . Thanks for help. 2011-10-23T04:37:29 *** Pawel__ has quit IRC (Quit: Page closed) 2011-10-23T04:38:08 cyphase: I have no manual to run it. In readme there are nothing. Can you give me a sample stdin line of parametrs to run? 2011-10-23T04:38:30 @tools 2011-10-23T04:38:31 Puj: tools could be Windows: http://aichallenge.org/tools.zip | Linux/MacOS X: http://aichallenge.org/tools.tar.bz2. 2011-10-23T04:38:33 aichallenge: McLeopold * r56a71d5 / website/submission.php : fix delete memcache for new org - http://git.io/glzSuQ 2011-10-23T04:38:44 Puj: suggested by me 2011-10-23T04:38:48 Do you have that Bel? 2011-10-23T04:38:51 Puj: i don't have list either 2011-10-23T04:38:58 Belerafon_L, i'm not sure, i haven't really gotten into any coding yet 2011-10-23T04:39:02 and don't think anyone have full list 2011-10-23T04:39:07 just create topic :) 2011-10-23T04:39:30 Ah, just jrc logs then :) 2011-10-23T04:39:38 5 top 10 players in a single game! http://aichallenge.org/visualizer.php?game=2856&user=66 2011-10-23T04:39:48 aichallenge: McLeopold * rbd9d121 / website/process_registration.php : fix delete memcache for new org - http://git.io/RwruPg 2011-10-23T04:39:52 All the parameters are listed in the specification 2011-10-23T04:40:24 ashoka: Yeah, just watched that one, interesting game, and GreenTea retakes the lead too. 2011-10-23T04:40:45 except the specification page is broken on the site at the moment :( 2011-10-23T04:40:47 three workers!! woohoo 2011-10-23T04:40:58 Then... Who programs ants on C++ language? Can anybody help me? 2011-10-23T04:41:21 *** xor__ has joined #aichallenge 2011-10-23T04:41:42 Belerafon_L, do you have @tools ? 2011-10-23T04:41:43 Hi there. 2011-10-23T04:41:56 Belerafon_L: why don't you run the sample scripts ? 2011-10-23T04:42:21 Is there a limit to the amount of processing a bot can do? 2011-10-23T04:42:27 they'll feed your program with the right input 2011-10-23T04:42:39 xor__, there's a time limit per turn 2011-10-23T04:42:43 xor__: yes, currently 500 milliseconds per turn 2011-10-23T04:42:47 I'm jealous they got such good pathfinding 2011-10-23T04:42:52 xor__: 500ms or more generally turntime given to the bot at the beginning of the game 2011-10-23T04:43:00 wow, 3 servers 2011-10-23T04:43:09 Puj: I have no pyton on my computer for tools. Can I work only with C++ sample or I need to use @tools and pyton? 2011-10-23T04:43:09 down to 17 now :( 2011-10-23T04:43:27 i can't believe those are all first - version bots 2011-10-23T04:43:28 Thank you. 2011-10-23T04:43:36 *** xor__ has quit IRC (Client Quit) 2011-10-23T04:43:36 *** mawif has quit IRC (Changing host) 2011-10-23T04:43:36 *** mawif has joined #aichallenge 2011-10-23T04:43:37 Belerafon_L: yes, install python and the tools 2011-10-23T04:43:43 500ms? i think it was 1000 2011-10-23T04:43:44 *** k333es has quit IRC (Quit: Page closed) 2011-10-23T04:43:45 Thanks 2011-10-23T04:43:45 bobbydroptable: they all participated in the beta 2011-10-23T04:43:52 *** Nakyuant has joined #aichallenge 2011-10-23T04:43:56 ahh... 2011-10-23T04:43:58 :( 2011-10-23T04:44:15 bobbydroptable: my bot is version #1 here, but it is the same as my version 10 from beta 2011-10-23T04:44:31 Belerafon_L, either install Python and use the tools or just dl them and copy the parameters you would like to use. The have good command line examples. But installing Python is relatively painless. 2011-10-23T04:44:33 *** amstan has quit IRC (Ping timeout: 245 seconds) 2011-10-23T04:44:59 bobbydroptable: but unless some people changed their name, I see new faces in the top 20 too. 2011-10-23T04:45:51 *** ztfw has joined #aichallenge 2011-10-23T04:46:07 bobbydroptable: betatesters 2011-10-23T04:46:12 *** avdg1 has joined #aichallenge 2011-10-23T04:46:18 all accounts were reseted 2011-10-23T04:46:23 *** welterde has quit IRC (Ping timeout: 260 seconds) 2011-10-23T04:47:28 bobbydroptable: but it is a long time before finals. in the last competition I didn't learn about it until the last three or four weeks and I was able to catch up and finish high. 2011-10-23T04:47:29 *** aboeing has joined #aichallenge 2011-10-23T04:47:35 only i cant see anything on terreble colored new vizualizer? 2011-10-23T04:47:37 *** volkhin has joined #aichallenge 2011-10-23T04:47:44 fair nuff 2011-10-23T04:47:52 i will get there ^^ 2011-10-23T04:47:57 *** Puj has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2011-10-23T04:48:23 does anyone have any sample C++ code for the challenge that actually preserves the state of objects between turns? I just realised the example code clears the state every turn.. 2011-10-23T04:48:50 here is a quick game: http://aichallenge.org/visualizer.php?game=2892 2011-10-23T04:49:41 wow 2011-10-23T04:49:43 hahaha 2011-10-23T04:50:08 what is the "Won at X" statistic? 2011-10-23T04:50:10 *** avdg has quit IRC (Ping timeout: 260 seconds) 2011-10-23T04:50:44 bobbydroptable, the turn in which the winner gained the lead 2011-10-23T04:51:02 oo 2011-10-23T04:51:03 okay 2011-10-23T04:51:43 ChrisH_: lol 2011-10-23T04:52:14 http://aichallenge.org/visualizer.php?game=2901 2011-10-23T04:52:55 *** Pawel__ has joined #aichallenge 2011-10-23T04:52:56 cyphase: I see the matrix in that game. 2011-10-23T04:54:21 4 starter packs? lol 2011-10-23T04:54:23 cyphase: Is it bubbles? 2011-10-23T04:54:32 lol 2011-10-23T04:55:09 hay - after my last post when i was receiving classNotDefined error setting up patgh didn't help. I just realized that the problem was probably in tutorial.sh (I changed "python tools" to "tools") but now i am having this error: Traceback (most recent call last): File "/home/pawel/TryChellenge/tools/engine.py", line 92, in run_game sandbox.start(bot_cmd) File "/home/pawel/TryChellenge/tools/sandbox.py", line 385, in star 2011-10-23T04:55:12 McLeopold: game rate is way up now with the new code and three workers playing a bunch of starter bot games 2011-10-23T04:55:27 wouldnt be better just to upload a code and see how it works? 2011-10-23T04:55:59 cyphase: awesome 2011-10-23T04:56:24 imagine at move 500, one of them starts expanding out and getting all that food 2011-10-23T04:56:43 aichallenge: McLeopold * r25ee4cc / ants/ants.py : fix negative ant location bug - http://git.io/vsr92g 2011-10-23T04:58:23 *** g0llum has joined #aichallenge 2011-10-23T04:58:35 do u know what to do? 2011-10-23T04:59:07 Traceback (most recent call last): File "/home/pawel/TryChellenge/tools/engine.py", line 92, in run_game sandbox.start(bot_cmd) File "/home/pawel/TryChellenge/tools/sandbox.py", line 385, in start raise SandboxError('Failed to start {0}'.format(shell_command)) SandboxErr 2011-10-23T04:59:39 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T05:01:48 *** Pawel__ has quit IRC (Quit: Page closed) 2011-10-23T05:02:03 How many games can I expect my bot to play in a day? 2011-10-23T05:02:08 On the server 2011-10-23T05:02:28 SharkMonkey: unknown because we haven't had a "normal" day yet 2011-10-23T05:02:49 aichallenge: McLeopold * r879b445 / website/mdx_latex.py : fix tutorial linux command issue - http://git.io/ZurCHA 2011-10-23T05:02:51 yesterday was swamped by tons of new accounts, and today there were server problems 2011-10-23T05:03:27 can anyone suggest exploration algorhitms? 2011-10-23T05:03:44 i want to map food to closest ants and rest go to explore for new food 2011-10-23T05:03:46 each turn 2011-10-23T05:04:09 ChrisH : Be happy today. tomorrow will be worse! (from Murphy's laws) 2011-10-23T05:04:32 ashoka: I'm just answering SharkMonkey's question 2011-10-23T05:04:57 Thanks for answering :) 2011-10-23T05:04:57 *** philxyz has joined #aichallenge 2011-10-23T05:05:35 I expected the high ranked players to have a lot more bot versions 2011-10-23T05:05:39 cirno_the_greate: if you know what parts of the map you haven't seen yet, and pick points evenly spread out in that area, that ought to help 2011-10-23T05:05:40 I'm making an Ada starter package - does the license need to be Apache 2.0? 2011-10-23T05:06:12 *** Belerafon_L has left #aichallenge 2011-10-23T05:06:24 sharkmonkey me too. most of them were beta players 2011-10-23T05:06:37 ChrisH_: good idea 2011-10-23T05:06:38 thanks :) 2011-10-23T05:06:44 *** avers_ has joined #aichallenge 2011-10-23T05:07:55 How long did the beta run for? 2011-10-23T05:08:29 bobbydroptable and SharkMonkey: I will also add that a lot of the top players competed in the previous ai challenge, so they have experience with some of the techniques. This makes it easier for them to zero in on their implementation without as much trial and error. 2011-10-23T05:08:46 this rugby game is becoming interesting 2011-10-23T05:08:52 ChrisH_: because just mapping ants to closest food funny and useless :) http://aichallenge.org/visualizer.php?game=2633&user=52 2011-10-23T05:08:55 SharkMonkey: the overall beta started in May, but the current form of the game only about 3 weeks ago 2011-10-23T05:09:36 synchronised ants! brilliant 2011-10-23T05:10:06 cirno_the_greate: That looks a lot like my bot did at version 3 about three or four weeks ago. :) 2011-10-23T05:10:16 *** klasbj has left #aichallenge 2011-10-23T05:10:23 Guys, does starter package code have to be released under Apache 2.0? (There's no linking exception for Adacore's compiler runtime library) 2011-10-23T05:10:25 mass suicide for Black at 210 2011-10-23T05:10:26 lol 2011-10-23T05:10:34 bobbydroptable: yep :) 2011-10-23T05:10:39 that's my collision detect, heh 2011-10-23T05:10:49 *** kronos has quit IRC (Ping timeout: 256 seconds) 2011-10-23T05:10:49 lol 2011-10-23T05:10:54 nice 2011-10-23T05:11:04 Is there a website up for the old ai challenges? 2011-10-23T05:11:25 every time they given order + clockwise backup variants when collision detected. 2011-10-23T05:11:32 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T05:11:35 like two ants want to step on 1 square 2011-10-23T05:11:38 Quick question: does the makefile that we upload get used as is by the game server when compiling the code? 2011-10-23T05:11:39 first going north 2011-10-23T05:11:40 SharkMonkey: not at the moment, but I think they want to get archives of them hosted somewhere. 2011-10-23T05:11:44 and second going east 2011-10-23T05:11:50 so i just deleted all my path finding code yay... start over 2011-10-23T05:11:51 after first claimed square 2011-10-23T05:11:51 *** noam has quit IRC (Read error: Connection reset by peer) 2011-10-23T05:11:59 second going to south then 2011-10-23T05:12:09 (on purpose) 2011-10-23T05:12:18 and they never stay on place 2011-10-23T05:12:25 sometimes they haven't any variants 2011-10-23T05:12:28 *** aboeing has quit IRC (Ping timeout: 265 seconds) 2011-10-23T05:12:29 and only kill themself 2011-10-23T05:12:30 :) 2011-10-23T05:12:49 need refactor collision map to two-pass check 2011-10-23T05:13:33 cirno_the_greate: that will help a lot 2011-10-23T05:13:50 *** philxyz has quit IRC (Quit: Page closed) 2011-10-23T05:14:47 *** Nakyuant has quit IRC (Quit: Page closed) 2011-10-23T05:14:48 is loadtime counted from when you receive "ready"? 2011-10-23T05:15:05 ah yea, it does say that 2011-10-23T05:15:44 ChrisH_: oh, you writing bot in Go 2011-10-23T05:15:49 mad props to you ;) 2011-10-23T05:16:14 ChrisH_: have you tried plan9 or inferno before? 2011-10-23T05:16:16 aichallenge: McLeopold * rc42cc5f / website/profile.php : update to next game display - http://git.io/nWOxZQ 2011-10-23T05:16:25 cirno_the_greate: thx this is my project to learn go with 2011-10-23T05:16:50 ChrisH_: same for me but for haskell :) 2011-10-23T05:16:52 although I already had some practice reimplementing most of my bot from planet wars as a trial 2011-10-23T05:16:54 man i don't believe i didn't think of this earlier ==" 2011-10-23T05:17:00 hopefully this works 2011-10-23T05:18:16 *** ashoka has quit IRC (Ping timeout: 265 seconds) 2011-10-23T05:18:18 well, I came in #2 in my latest game, but good enough to move me up a couple ranks to #7. that won't hold though 2011-10-23T05:18:46 yikes i've dropped to 20 since the servers gone up 2011-10-23T05:18:46 lol 2011-10-23T05:18:55 gotta fix this bot fast 2011-10-23T05:18:57 *** avers_ has joined #aichallenge 2011-10-23T05:19:06 http://aichallenge.org/visualizer.php?game=2809&user=1962 mass-starter kit suicide 2011-10-23T05:19:24 at last turn 2011-10-23T05:19:32 simultaneously :D 2011-10-23T05:19:32 bobbydroptable: its not your rank now that matters... its what happens after the real comp begines 2011-10-23T05:19:42 *** choas has joined #aichallenge 2011-10-23T05:19:43 what you mean the real comp 2011-10-23T05:19:58 isn't this the comp? 2011-10-23T05:20:00 well the final ranking are deteriment at the end 2011-10-23T05:20:08 *** AlvinMinring_ has joined #aichallenge 2011-10-23T05:20:15 ah 2011-10-23T05:20:16 bobbydroptable: read the bottom of the home page: 2011-10-23T05:20:18 submissions are closed then the bots battle it out 2011-10-23T05:20:21 aichallenge: McLeopold * rd71bb71 / website/profile.php : update to next game display - http://git.io/msPAmA 2011-10-23T05:20:26 there is a lot of black horses here :p 2011-10-23T05:20:39 what do the winners get? 2011-10-23T05:20:44 loo 2011-10-23T05:20:46 lol 2011-10-23T05:20:49 internship at google? blizzard? 2011-10-23T05:20:52 XD 2011-10-23T05:20:59 McLeopold: $logged_in is already that check no? 2011-10-23T05:21:01 does anyone here know how makefiles are dealth with by the server? 2011-10-23T05:21:02 bobbydroptable: awesome line in CV, doesn't it? 2011-10-23T05:21:09 AlvinMinring_: ignored 2011-10-23T05:21:10 *** AlvinMinring has quit IRC (Ping timeout: 265 seconds) 2011-10-23T05:21:21 AlvinMinring_: https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py 2011-10-23T05:21:25 hehe 2011-10-23T05:21:29 AlvinMinring_: no makefiles 2011-10-23T05:22:10 well, time for sleep, good luck all 2011-10-23T05:22:30 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-23T05:22:32 cirno_the_greate: thx I'll look into that 2011-10-23T05:22:32 do we get some kind of 'official' certificate or something? 2011-10-23T05:22:33 lol 2011-10-23T05:23:12 cyphase: you want certificate from computer club? 2011-10-23T05:23:21 this competition held not by company or whatever 2011-10-23T05:24:14 *** Pawel__ has joined #aichallenge 2011-10-23T05:24:18 ok, I see, the set of compiler options is predetermined in the server script 2011-10-23T05:24:31 that's too bad 2011-10-23T05:24:41 AlvinMinring_: yep. And you can open issue at github, if you need something specific 2011-10-23T05:24:50 what's your problem? 2011-10-23T05:24:57 janzert: I just made it so you can only see your own wait time 2011-10-23T05:25:06 I wanted to use gcc's -std=c++0x 2011-10-23T05:25:06 it already was though 2011-10-23T05:25:12 *** ChrisH_ has joined #aichallenge 2011-10-23T05:25:23 *** Sunshowers has left #aichallenge 2011-10-23T05:25:24 $logged_in is only set if it is the current user or an admin viewing the page 2011-10-23T05:25:30 Hello Guys!! I just submitted a new version of my Bot - do You know if I should do anything else - because now I can see both versions on my profile - but only the older one is playing games for now?? 2011-10-23T05:25:31 oh 2011-10-23T05:25:35 oh well 2011-10-23T05:25:58 McLeopold: I just got the "other" org filtered version of the standings page. Do you want to see the page source? 2011-10-23T05:26:05 yes 2011-10-23T05:26:10 and url 2011-10-23T05:26:57 any idea? 2011-10-23T05:27:07 McLeopold: http://pastebin.com/6F4pKhSK 2011-10-23T05:27:25 McLeopold: URL is http://aichallenge.org/rankings.php?page=1 2011-10-23T05:27:43 Pawel__: your new submission just hasn't had a chance to play yet 2011-10-23T05:27:55 AlvinMinring_: oh. Open issue at github than. I'm not c++ guy anyway (: 2011-10-23T05:28:01 next game you get it should be the new one 2011-10-23T05:28:05 *** noam has joined #aichallenge 2011-10-23T05:28:50 *** Dispatch has joined #aichallenge 2011-10-23T05:29:07 McLeopold: curiously, I'm pretty sure that page showed up after I clicked back on my browser 2011-10-23T05:29:14 aha - thanks! I cannot test my bots on my local machine though - is there anyone else with similar problem? Had u problem like this janzert? 2011-10-23T05:29:16 what kind of processing power do the bots have? 2011-10-23T05:29:37 *** marijnfs has quit IRC (Quit: Lost terminal) 2011-10-23T05:29:42 no, but I don't have my own bot ;) 2011-10-23T05:29:56 McLeopold: yeah, I had visited page 17, then page 16, then back twice and got the filtered page 1 2011-10-23T05:30:13 aichallenge: McLeopold * rc3763ee / website/sql.php : fix for next game in logic - http://git.io/BY4AMQ 2011-10-23T05:30:24 *** JamesMG has quit IRC (Read error: Operation timed out) 2011-10-23T05:31:02 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T05:31:24 *** kronos has joined #aichallenge 2011-10-23T05:31:29 *** awishformore has joined #aichallenge 2011-10-23T05:32:11 cirno_the_greate: Ok thanks, I'll try that if it turns out I really need it. 2011-10-23T05:32:26 *** rr5 has joined #aichallenge 2011-10-23T05:32:33 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-23T05:32:57 cyphase: 1 gig of ram and some cpu cycles 2011-10-23T05:33:00 @ranking 2011-10-23T05:33:01 McLeopold: No! 2011-10-23T05:33:08 @rankings 2011-10-23T05:33:14 McLeopold: Top 10 players: GreenTea(73.7), FlagCapper(72.3), Ender(70.5), sir_macelon(69.3), MomoBot(69.0), ThisIsNotABug(68.2), ChrisH(68.1), ZumZoom(67.8), olex.s(67.1), beanhome(66.8) 2011-10-23T05:34:37 *** Guest007 has joined #aichallenge 2011-10-23T05:35:55 *** ashoka has joined #aichallenge 2011-10-23T05:36:14 *** Pawel__ has quit IRC (Quit: Page closed) 2011-10-23T05:37:16 @help 2011-10-23T05:37:16 cirno_the_greate: (help [] []) -- This command gives a useful description of what does. is only necessary if the command is in more than one plugin. 2011-10-23T05:37:21 @list 2011-10-23T05:37:21 cirno_the_greate: AIChallenge, Admin, Alias, Anonymous, BadWords, Channel, ChannelLogger, ChannelStats, Config, Ctcp, Dict, Dunno, Factoids, Filter, Format, Games, Google, Herald, Internet, Later, Limiter, Math, Misc, Network, News, Note, Owner, Plugin, Quote, QuoteGrabs, RSS, Reply, Scheduler, Seen, Services, ShrinkUrl, Status, String, Success, Time, Todo, Topic, URL, Unix, User, Utilities, and (1 more message) 2011-10-23T05:37:41 *** kronos has quit IRC (Quit: leaving) 2011-10-23T05:37:58 *** JamesMG has joined #aichallenge 2011-10-23T05:38:21 *** avers_ has joined #aichallenge 2011-10-23T05:38:57 cirno_the_greate: are you still working on the starter package? 2011-10-23T05:39:26 cirno_the_greate: the Haskell starter package, that is 2011-10-23T05:39:34 *** n4nd0 has joined #aichallenge 2011-10-23T05:39:34 boegel: already done 2011-10-23T05:39:56 boegel: https://github.com/qacek/ants-haskell/pull/1 2011-10-23T05:40:09 go to ranking page 16, sort by recent, almost everyone who has no games was a resubmitter 2011-10-23T05:40:26 boegel: and here - https://github.com/aichallenge/aichallenge/issues/320 2011-10-23T05:40:46 McLeopold: that's a good thing right? 2011-10-23T05:41:25 craklyn wins the prize for most dedicated with 60 versions :) http://aichallenge.org/profile.php?user=1231 2011-10-23T05:42:23 actually, he looks like he was coding inbetween each submission 2011-10-23T05:42:54 cirno_the_greate: did you also implement all the helper functions? 2011-10-23T05:42:57 janzert: it means the per user is working correctly 2011-10-23T05:43:05 * janzert nods 2011-10-23T05:43:12 cirno_the_greate: I didn't see a "occupied" or "unoccupied" function for example 2011-10-23T05:43:12 Oh, the column sort is javascript huh. Will only sort the rankings on the current page? 2011-10-23T05:43:21 yep 2011-10-23T05:43:50 janzert: we should go to bed X( 2011-10-23T05:44:03 boegel: mmm 2011-10-23T05:44:08 yea, few hours ago :P 2011-10-23T05:44:09 you could just use isElem for point 2011-10-23T05:44:14 i'll add it right now 2011-10-23T05:44:21 janzert: let me inspire you... 2011-10-23T05:44:24 *** McLeopold has quit IRC (Read error: Connection reset by peer) 2011-10-23T05:44:45 cirno_the_greate: ok, I should probably fork your git repo then :) 2011-10-23T05:45:04 cirno_the_greate: because it seems we're working beside each other on the same things... 2011-10-23T05:45:19 boegel: okay 2011-10-23T05:45:26 cirno_the_greate: also, we need to get the last version of the starter package on the website (not just in the main git repo) 2011-10-23T05:45:36 i planned to use qacek's repo and pull all changes to his repo 2011-10-23T05:46:04 boegel: zip files are in aichallenge git repo? 2011-10-23T05:46:18 Did the older AI challenges have rewards? 2011-10-23T05:46:55 cirno_the_greate: no, I don't think so, but we need someone to update the ZIP file 2011-10-23T05:47:49 SharkMonkey: excellent record in CV perhaps 2011-10-23T05:48:23 boegel: okay, which functions are needed? 2011-10-23T05:48:26 *** Harpyon has joined #aichallenge 2011-10-23T05:48:35 isOccupied, what else? 2011-10-23T05:48:50 It probably looked better on a CV when it was the Google AI challenge 2011-10-23T05:48:52 nothing comes to my mind at the moment 2011-10-23T05:50:06 *** helkafen has joined #aichallenge 2011-10-23T05:50:54 Will submitting a new version reset my place in the queue and put me at the back of the line? 2011-10-23T05:51:58 SharkMonkey: nope 2011-10-23T05:52:03 i think it will forward instead 2011-10-23T05:52:06 but will reset rank 2011-10-23T05:54:53 *** pistacchio has joined #aichallenge 2011-10-23T05:56:49 this might be a really stupid question as i am getting started 2011-10-23T05:57:00 but why are all the radii passed as squared values? 2011-10-23T05:58:19 when you calculate a distance, just do dx^2 + dy^2 to get a distance squared 2011-10-23T05:58:43 awishformore, When you calculate the distance between two points you do sqrt((x1-x2)^2 + (y1-y2)^2). Instead you can skip doing the square root and just campare the value against the squared radius 2011-10-23T05:58:48 saves having to do a sqrt() which can be computationally expensive 2011-10-23T05:59:13 what's the dotted line around ant hills in the visualizer? 2011-10-23T05:59:31 seems like something slow languages would need :D 2011-10-23T05:59:33 indicates color of owner 2011-10-23T05:59:38 but ok, that's sensible ;) 2011-10-23T06:01:19 *** irchs has joined #aichallenge 2011-10-23T06:03:47 *** jtamer has joined #aichallenge 2011-10-23T06:04:48 *** petterw has quit IRC (Changing host) 2011-10-23T06:04:48 *** petterw has joined #aichallenge 2011-10-23T06:09:06 cirno_the_greate: I'd name them 'occupied', 'unoccupied' and also the 'myHills', 'enemyHills', 'myAnts', etc functions 2011-10-23T06:09:50 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T06:10:25 *** Puj has joined #aichallenge 2011-10-23T06:11:01 boegel: okay 2011-10-23T06:11:27 Sunday breakfast at its finest :) - http://twitpic.com/74kv0m 2011-10-23T06:11:28 cirno_the_greate: see the bottom of page https://github.com/aichallenge/aichallenge/wiki/Ants-Starter-Pack-Guide 2011-10-23T06:11:36 cirno_the_greate: I think all of those might come in handy 2011-10-23T06:12:13 no any good ideas 2011-10-23T06:12:17 *** SofiaWhite has quit IRC (Quit: Page closed) 2011-10-23T06:12:20 boegel: very good 2011-10-23T06:12:21 thanks 2011-10-23T06:12:25 3 day in depression 2011-10-23T06:13:31 Accoun: look for food 2011-10-23T06:13:32 *** Belerafon_L has joined #aichallenge 2011-10-23T06:13:34 and raze hills 2011-10-23T06:14:15 Accoun, rewrite your Bot. That usually helps finding new ways. 2011-10-23T06:14:18 Accoun, make small iterative 2011-10-23T06:14:20 progress 2011-10-23T06:15:14 mmm 2011-10-23T06:15:20 that was wonderful breakfast 2011-10-23T06:15:20 cirno_the_greate: I might have some time to help out today... just let me know what you'll look into 2011-10-23T06:15:37 boegel: this isn't hard really 2011-10-23T06:15:48 have you started with collision detect? 2011-10-23T06:15:55 i would like your implementation details 2011-10-23T06:16:06 i made mine, but i'm completely dissapoint with it 2011-10-23T06:16:13 not algorhitm, implementation 2011-10-23T06:16:15 Accoun : depression is not necessarily a bad thing in this context, usually good ideas and progress follow it. :-) 2011-10-23T06:16:28 because first of all it's impure ;( 2011-10-23T06:16:43 just hooked up IORef with list of Point to GameState 2011-10-23T06:17:01 *** avers_ has joined #aichallenge 2011-10-23T06:17:08 and forced some functions to run inside IO monad 2011-10-23T06:17:39 Every time I look at my profile the next game seems forther off 2011-10-23T06:17:56 Accoun : BTW, is that TCP server yours ? a big thank. maps there need to be updated though. Think janzert has removed some of buggy maps from the competition. 2011-10-23T06:19:29 I can't run my bot on @tools. I use C++ and replaced first standart bot in play_one_game.cmd to "my.exe". Script run without errors but there are no input stream to my.exe. Input function for istream is not called. Where the problem? 2011-10-23T06:20:12 cirno_the_greate: no, I've only submitted the starter bot, haven't done any work on my own bot yet 2011-10-23T06:21:08 *** UncleVasya has joined #aichallenge 2011-10-23T06:22:05 Belerafon_L, You mean you get no input to stdin in your program? 2011-10-23T06:22:25 yes, from tools 2011-10-23T06:22:55 On terminal input stdin works 2011-10-23T06:23:47 Belerafon_L, Have you tried doing the same thing with play_one_game_live? 2011-10-23T06:24:54 Puj: On play_one_game_live no work too 2011-10-23T06:26:02 Belerafon_L, And you are using the starter bot for C++ unchanged? 2011-10-23T06:26:09 *** Redgis has joined #aichallenge 2011-10-23T06:26:29 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T06:26:37 I write in play_one_game_live: .....aze_04p_01.map" %* "my.exe" "python ""%~dp0sample_bots\python\LeftyBot.py""" ..... 2011-10-23T06:26:46 it is correct? 2011-10-23T06:27:35 yes, original starter bot 2011-10-23T06:27:38 *** rr5 has quit IRC (Remote host closed the connection) 2011-10-23T06:27:59 Belerafon_L, it seems you may have to do "../My.exe" since you are in your tools dir. 2011-10-23T06:28:20 *** Puj has quit IRC (Remote host closed the connection) 2011-10-23T06:28:31 *** Puj has joined #aichallenge 2011-10-23T06:28:35 Belerafon_L: play_one_game has a problems with launching bots from subfolders if it not fixed yet. 2011-10-23T06:28:41 *** KODer has joined #aichallenge 2011-10-23T06:29:04 Belerafon_L: try to launch a game with few my.exe - s. 2011-10-23T06:29:11 (Puj) Belerafon_L, it seems you may have to do "../My.exe" since you are in your tools dir. -- Sorry on train. 2011-10-23T06:29:22 *** Guest86160 has quit IRC (Quit: Page closed) 2011-10-23T06:29:31 *** KODer has quit IRC (Client Quit) 2011-10-23T06:29:52 I put "my.exe" to "tools", without subfolsers. Is it correct? 2011-10-23T06:30:13 Belerafon_L, should work 2011-10-23T06:30:27 boegel: you still here? 2011-10-23T06:30:52 Is my.exe in the same directory with play_one_game 2011-10-23T06:30:57 * ? 2011-10-23T06:31:38 UncleVasya, yes, same. my.exe runs from tools, work, but have no input stream 2011-10-23T06:31:50 i doesn't see any meaning in passable/unoccupied functions for starter pack. First of all because ants doesn't have any state to memorize orders 2011-10-23T06:31:59 Belerafon_L: or you can try to place sample bots from subfolders to the directory with play_one_game . 2011-10-23T06:32:19 Belerafon_L: And you running only copies of your bot? 2011-10-23T06:32:46 boegel: i can make functions to determine hill status (occupied or not) 2011-10-23T06:33:02 *** avers_ has joined #aichallenge 2011-10-23T06:33:26 boegel: write to me on github when you will be back or ping me if i'll be available here 2011-10-23T06:33:50 UncleVasya: no, I run my bot and 3 sample bots 2011-10-23T06:33:54 Like .aze_04p_01.map" %* my.exe my.exe 2011-10-23T06:34:47 Belerafon_L: Sample bots are in the subfolder, that can be an issue. 2011-10-23T06:34:56 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2011-10-23T06:35:59 I could not remember how I have fixed this but I remember I had this problem: if any bot is not from the directory with play_one_game 2011-10-23T06:36:02 Belerafon_L, it seems my samples bots are working out of subfolders and mybot out of ../. is that odd? 2011-10-23T06:36:34 *** george has quit IRC (Ping timeout: 265 seconds) 2011-10-23T06:36:39 *** pedrosorio has joined #aichallenge 2011-10-23T06:36:39 UncleVasya, rather 2011-10-23T06:38:30 *** dangoodman has quit IRC (Ping timeout: 265 seconds) 2011-10-23T06:39:09 I try move GreedyBot.py from 'sample_bots\python" to "sample_bots", and it crash. My bee I need to move some other files? 2011-10-23T06:39:28 *** ashoka has quit IRC (Ping timeout: 265 seconds) 2011-10-23T06:39:42 GreedyBot very slow and timeouts 2011-10-23T06:39:55 (not related) 2011-10-23T06:40:00 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T06:40:13 *** HaraKiri has joined #aichallenge 2011-10-23T06:40:13 Belerafon_L: Why to sample_bots/ ? Directly to the folder with play_one_game.cmd, place it near it. 2011-10-23T06:40:24 *** irchs has quit IRC (Quit: irchs) 2011-10-23T06:41:21 Belerafon_L, try doing test_bot or play one game with only your bot and -fill 2011-10-23T06:41:55 Just ry to launch a game with only your bots like ...aze_04p_01.map" %* my.exe my.exe, if the problem still will be there I cannot help :) 2011-10-23T06:41:58 And... How to modify "python ""%~dp0sample_bots\GreedyBot.py""" ? to "python GreedyBot.py" ? I do not know pyton syntax 2011-10-23T06:42:56 *** levis501 has quit IRC (Quit: levis501) 2011-10-23T06:43:01 *** skyramp has joined #aichallenge 2011-10-23T06:43:15 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2011-10-23T06:44:28 Belerafon_L, "python ""%~dp0GreedyBot.py""" 2011-10-23T06:46:59 *** SharkMonkey has quit IRC (Quit: leaving) 2011-10-23T06:47:17 *** avers_ has joined #aichallenge 2011-10-23T06:47:30 *** olexs has joined #aichallenge 2011-10-23T06:47:31 *** xathis has joined #aichallenge 2011-10-23T06:47:34 Puj: it crash... 2011-10-23T06:48:19 *** nishant has joined #aichallenge 2011-10-23T06:48:27 Did you try what uncle said only giving two bots as input. Both are your bots my.exempt 2011-10-23T06:48:28 http://ash.webfactional.com/replay.610 <--- proof-of-concept ant-graffiti 2011-10-23T06:48:49 sry gtg gl 2011-10-23T06:49:55 lol Alexer 2011-10-23T06:50:07 *** nogood_cz has joined #aichallenge 2011-10-23T06:50:47 *** skyramp has quit IRC (Quit: Page closed) 2011-10-23T06:51:02 Puj: I try this http://pastebin.com/6T41zJBh and this - http://pastebin.com/pfU8qxdN is't work 2011-10-23T06:51:57 *** nishant_ has joined #aichallenge 2011-10-23T06:52:48 anybody there? 2011-10-23T06:52:48 *** nishant has quit IRC (Ping timeout: 240 seconds) 2011-10-23T06:53:04 *** Puj has quit IRC (Ping timeout: 244 seconds) 2011-10-23T06:53:04 hi 2011-10-23T06:53:09 hello 2011-10-23T06:53:36 nishant_: ? 2011-10-23T06:53:59 what exactly is the purpose of this channel? 2011-10-23T06:54:09 Belerafon_L: try putting the executable into the tools dir and not using the entire path, that works for me, even though it's annoying why the script would be so ill-programmed 2011-10-23T06:54:28 oh wait 2011-10-23T06:54:31 never mind 2011-10-23T06:54:37 different topic 2011-10-23T06:56:42 for people to chat exchange ideas help others with tools report bugs 2011-10-23T06:58:19 *** vladimirfol has quit IRC (Ping timeout: 265 seconds) 2011-10-23T06:59:54 *** UncleVasya has joined #aichallenge 2011-10-23T07:01:09 *** nishant_ has quit IRC (Ping timeout: 255 seconds) 2011-10-23T07:01:09 is there any way to print out (debug) information in python? "print"-ing creates errors, as does file writing 2011-10-23T07:01:21 print to stderr 2011-10-23T07:01:24 *** kronos has joined #aichallenge 2011-10-23T07:01:30 stdout used for communication with server 2011-10-23T07:01:44 and launch playgame.py with -Ee option 2011-10-23T07:01:54 and look for logs in game_logs folder 2011-10-23T07:02:50 i put a simple "sys.stderr.write("test")" it it, still crashes even with "-Ee" 2011-10-23T07:03:01 *** george has joined #aichallenge 2011-10-23T07:04:09 it crashes exactly at sys.stderr.write? 2011-10-23T07:04:46 *** flo_ has joined #aichallenge 2011-10-23T07:06:07 yes, well the error is in the following line: "IndentationError: unindent does not match...". commenting out the stderr-line fixes problem 2011-10-23T07:06:38 well, error says that there is problem in indentation 2011-10-23T07:06:41 not stderr 2011-10-23T07:06:42 there must be an indentation error on the commented out line, then 2011-10-23T07:07:29 i know it should be that easy, but it really isnt that. indentation isnt THAT hard :x. also, line works fine in other programs 2011-10-23T07:08:07 then your python broken 2011-10-23T07:08:11 seems so 2011-10-23T07:08:12 there isn't many variants 2011-10-23T07:08:19 I hope you're using a sensible python editor mode that doesn't mix leading tabs and indentation on non-continuation lines 2011-10-23T07:08:25 just blame on interpretator 2011-10-23T07:08:27 tabs and spaces* 2011-10-23T07:09:36 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-23T07:10:12 *** danielharan has joined #aichallenge 2011-10-23T07:10:42 oh wow got it, thanks 2011-10-23T07:10:57 yarr 2011-10-23T07:11:46 output should be in .error files right? they're there, but empty :< 2011-10-23T07:12:31 *** GreenTea has joined #aichallenge 2011-10-23T07:12:35 probably need a command line arg somewhere to enable that file 2011-10-23T07:12:42 or writing to it 2011-10-23T07:13:05 check the shell/cmd files and the python game runner script they invoke 2011-10-23T07:13:12 *** danielha_ has joined #aichallenge 2011-10-23T07:13:18 i used the "-Ee" cirno_the_greate told me, and also tried "--log_error --log_stderr" which i got from forums 2011-10-23T07:13:18 under tools 2011-10-23T07:13:21 *** danielha_ has quit IRC (Remote host closed the connection) 2011-10-23T07:13:28 go to the sauce 2011-10-23T07:13:33 ok will do 2011-10-23T07:13:33 *** danielharan has quit IRC (Read error: Connection reset by peer) 2011-10-23T07:13:35 the sauce is great :D 2011-10-23T07:13:39 Hi 2011-10-23T07:13:43 Heyo 2011-10-23T07:13:51 *** danielharan has joined #aichallenge 2011-10-23T07:14:05 python sauce 2011-10-23T07:14:07 anybody from top want to make firght on tcp server? 2011-10-23T07:14:13 *fight 2011-10-23T07:14:24 GreenTea: everyone is too scared of you 2011-10-23T07:14:32 )) 2011-10-23T07:14:35 *** Savaron has joined #aichallenge 2011-10-23T07:14:52 I'll fight you once my ants aren't retarded 2011-10-23T07:14:57 until then, you just wait, mister 2011-10-23T07:15:06 must admit you seem to adapt to the new rules very quickly GreenTea 2011-10-23T07:15:25 GreenTea: which tcp are you on? 2011-10-23T07:15:32 AndrewBC: master... GreanTea is Master GreanTea ;) 2011-10-23T07:15:36 *** nogood_cz has quit IRC (Quit: Page closed) 2011-10-23T07:15:48 *** ejls has quit IRC (Ping timeout: 245 seconds) 2011-10-23T07:15:49 Pff 2011-10-23T07:15:53 *gong* 2011-10-23T07:15:53 there's 2 months yet 2011-10-23T07:16:11 Way to soon to bend my knee and kiss any asses 2011-10-23T07:16:22 and Winter is Coming. 2011-10-23T07:16:24 lol 2011-10-23T07:16:42 *** ejls has joined #aichallenge 2011-10-23T07:16:48 well GreenTea didn't too badly last time IIRC 2011-10-23T07:16:52 (been reading so much ASoIaF that I'm starting to talk like it :|) 2011-10-23T07:17:06 ASoIaF? 2011-10-23T07:17:08 delt0r_, new rules? 2011-10-23T07:17:16 A Song of Ice and Fire 2011-10-23T07:17:28 the game of thrones book series 2011-10-23T07:17:30 "oooh" 2011-10-23T07:17:37 pretty good fantasy series for th- ^ what he said 2011-10-23T07:17:41 Alexer: the rules in the beta where change quite drastically about 2-3 weeks ago 2011-10-23T07:17:42 olexs, maby on http://ash.webfactional.com, aim not play on it yet. 2011-10-23T07:18:14 delt0r_, ah.... 2011-10-23T07:18:18 oh right books from games --interesting. I tried to read a book based on Eve Online. It kinda sucked 2011-10-23T07:18:30 delt0r_, umm 2011-10-23T07:18:40 GreenTea: I'll begin working on my bot again next week, it's unmodified since the hills were introduced in the beta right now... I'll play you then :) 2011-10-23T07:18:41 delt0r_, gae of thrones is a tv series based on the books 2011-10-23T07:18:44 game* 2011-10-23T07:18:48 oh 2011-10-23T07:18:49 nah, Game of Thrones is the first book of the series, and also the name of the tv show 2011-10-23T07:19:09 which is also good, mind you 2011-10-23T07:19:14 GreenTea, you can just let your bot running indefinitely. at some point it'll bump into a good bot, hopefully :P 2011-10-23T07:19:42 if people just remember to advertise the tcp servers occasionally >:P 2011-10-23T07:19:57 olexs, ok) 7 place for unmidified bot its very good 2011-10-23T07:20:14 *** onensora has joined #aichallenge 2011-10-23T07:20:39 So, naive question here: how stable can we assume the rules to be? Like, can we assume that the view radius is greater than the attack radius? Are these values going to get determined once and for all some time down the road, or could they vary from game to game even in the latest stages of the challenge? 2011-10-23T07:20:51 GreenTea: the last mod I did was going for a hill kill when it's safe. there is no defence and no pathfinding optimizations for mazes at all... but it does better than I expected 2011-10-23T07:21:11 AlliedEnvy_: if its a parameter that you get passed you should assume that it can be somewhat fluid 2011-10-23T07:21:37 but basic constraints like view distance being grater than kill distance is a safe bet 2011-10-23T07:21:40 *** AlliedEnvy_ is now known as AlliedEnvy 2011-10-23T07:21:53 can't think why that would ever be different. 2011-10-23T07:22:07 AlvinMinring_: the values were stabilized in the beta, and most probably will not change until the challenge is over 2011-10-23T07:22:12 olexs, cool, I make special algorithm for exloration of mazes. but it is still not perfect.. 2011-10-23T07:22:23 ok, interesting, thanks 2011-10-23T07:22:27 AlliedEnvy: the reason is last time around there really needed to be some tweaking in the parameters. Too many hard coded stuff based on that 2011-10-23T07:22:32 GreenTea: I saw you have tried lisp :) 2011-10-23T07:22:48 Can anybody give me executable working version of ants bot for windows? I try on my @tools... 2011-10-23T07:23:12 * ñ++ 2011-10-23T07:23:18 exe file 2011-10-23T07:23:24 In fact i think GreenTea was one person who got burnt with the game turn limit change? 2011-10-23T07:23:48 *** samnon has joined #aichallenge 2011-10-23T07:23:54 Hey 2011-10-23T07:24:04 UncleVasya, yea) but in my crooked hands it is too slow. path finding was 6 time slower than on java... 2011-10-23T07:24:10 *** foRei has joined #aichallenge 2011-10-23T07:24:34 GreenTea: really? 2011-10-23T07:24:41 in lisp? 2011-10-23T07:24:44 delt0r_: Yep I see, but one could imagine making these basic values change dramatically to see how robust bots are. That would make creating a competent bot a whole lot more complicated though. 2011-10-23T07:24:52 with declared types? 2011-10-23T07:24:58 * delt0r_ is surprised 2011-10-23T07:25:04 with no declared tipes 2011-10-23T07:25:13 it is too complicated for me) 2011-10-23T07:25:16 *** thestinger has joined #aichallenge 2011-10-23T07:25:48 AlvinMinring_: yes. but its likely to be map based only.. ie turn limit for big maps and many players may be larger 2011-10-23T07:26:05 i see some code with declared tipes, and cant understant it. it seem very low level.. 2011-10-23T07:26:06 i expect that view radius to be constant over all maps.. 2011-10-23T07:26:14 like assembler... 2011-10-23T07:26:18 GreenTea: right... its not so hard 2011-10-23T07:26:22 delt0r_: yep, makes sense 2011-10-23T07:26:39 is also "meta information" in that lisp does not need declared types 2011-10-23T07:27:05 but then the JVM has had a *lot* of work put into it by some big players 2011-10-23T07:27:30 delt0r_, do you know some good book or tutoralal about declared types in sbcl? 2011-10-23T07:27:44 Java is fast enough if one knows his way around using proper data structures and algorithms correctly 2011-10-23T07:27:59 GreenTea: not really. Sort of learnt what i know from doing.. but i have not had a chance to do much lisp 2011-10-23T07:28:14 i just know that performance is not consider bad genrally 2011-10-23T07:28:24 and 6x slower than java is bad 2011-10-23T07:28:46 java is not slow... but its not SSE/MMX optimised C either 2011-10-23T07:29:37 GreenTea: what i have done in the past is getting everything working. Then add declarations where needed after that. 2011-10-23T07:30:04 olexs: true 2011-10-23T07:30:26 true in general for most langs... you need to understand what works with your tool... 2011-10-23T07:31:00 it's important to avoid O(n) and higher complexity stuff where possible, and not to overload GC with too many short-lived instances of objects 2011-10-23T07:31:01 olexs: my day job is replacing a C app with a java one. My current java version is more than 7 times faster 2011-10-23T07:31:14 that is not true anymore 2011-10-23T07:31:30 short lived objects are cheap with generational GC 2011-10-23T07:32:10 object pools typically are slower than not using them (in the case of pure java objects JDBC connection are a different story) 2011-10-23T07:32:33 delt0r_: my testing proved different... maybe those were large objects, but in my battle simulation search tree, it is much cheaper to perform all simulation on a single node instance, then rolling back changes, than creating a new instance for every node and discarding the non-optimal paths later 2011-10-23T07:33:13 well if you have profiled then sure... i belive you... 2011-10-23T07:33:27 but be warned on some JVMs it could well be slower 2011-10-23T07:33:50 also that is not really pooling 2011-10-23T07:33:57 true... I'd actually need to test it on the same JVM the server uses. by the way, what is it? :) 2011-10-23T07:34:12 pooling has a bad habit of polluting the cache 2011-10-23T07:34:21 OpenJDK 2011-10-23T07:34:27 ok, I can test that 2011-10-23T07:34:36 will be different from jdk 7 2011-10-23T07:34:44 more like a jdk 1.6.0 2011-10-23T07:34:47 i guess 2011-10-23T07:34:54 its is different that is for sure 2011-10-23T07:34:56 I'm still on JDK 6, winx64 here 2011-10-23T07:35:16 then i expect you are getting a pretty good indication of server side performance 2011-10-23T07:35:22 but never hurts to test 2011-10-23T07:35:37 unfortunately my bot works very close to time limit 1sec. therefore even small slowness is important for me. And i cant risk to rewrite all in lisp and then find out that i have timeouts everywhere... 2011-10-23T07:35:51 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-23T07:35:53 well it was fairly consistent, and I could add 2 plies of base search depth after that optimization without having timeouts on theserver 2011-10-23T07:36:03 GreenTea: yea.. what is your most experiance lang? 2011-10-23T07:36:18 java 2011-10-23T07:36:22 olexs: good to know 2011-10-23T07:36:39 GreenTea: its why i use it a lot 2011-10-23T07:36:45 i just know it well 2011-10-23T07:36:55 ) 2011-10-23T07:37:06 so easier for me to write efficient code in it 2011-10-23T07:38:10 GreenTea: are you doing big searches? 2011-10-23T07:38:36 and is your timing dynamic? ie put it on faster hardware it will use it? 2011-10-23T07:39:24 delt0r_, most time consuming is path finding and battle calculation 2011-10-23T07:39:48 olexs: also in your case with your nodes these are not short lived objects... so replacing objects that get into the next generation is more like the old java gc model 2011-10-23T07:39:55 GreenTea: Really? 2011-10-23T07:40:23 if time is near over then bot can skip some calculation and make more simple moves 2011-10-23T07:40:24 i guess the maps are quite big... but would think A* should handle it fine 2011-10-23T07:40:45 so on better hardwere it should play stronger 2011-10-23T07:40:52 right 2011-10-23T07:40:56 GreenTea: by battle calculation you mean the resolution? that is pretty fast for me, by far not the slowest element 2011-10-23T07:42:22 olexs, i try to find optimal solution in battle resolution 2011-10-23T07:42:27 therefore it is slow 2011-10-23T07:42:48 GreenTea: me too. the tree generation and simulation in total is slow, yes 2011-10-23T07:43:15 but then, it's the main "brain" of the bot, so it's allowed to be :) 2011-10-23T07:43:46 GreenTea: i though your bot was doing from watching a few games 2011-10-23T07:43:50 thought 2011-10-23T07:44:24 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T07:47:24 i will play few games on http://ash.webfactional.com/, if someone want to stop me from taking first plase - you are welcome :) 2011-10-23T07:48:28 on it :) 2011-10-23T07:50:15 I'll probably fail though, #1 vs #7 should be pretty obvious... but no matter, more games for analysis is always better 2011-10-23T07:51:26 *** avers_ has joined #aichallenge 2011-10-23T07:53:28 *** IhateIt has joined #aichallenge 2011-10-23T07:53:42 Hello 2011-10-23T07:54:37 *** FusionX has quit IRC (Ping timeout: 240 seconds) 2011-10-23T07:55:08 Can anyone help me?? I don't understand how this bot should work? I keep trying and I changed only a little in main method form starter package (java) and it keeps crushing always when I try to test it. WHY??? 2011-10-23T07:55:47 *** den has joined #aichallenge 2011-10-23T07:56:38 *** den has left #aichallenge 2011-10-23T07:56:39 can anyone help me? 2011-10-23T07:56:48 *** bobbydroptable has quit IRC (Ping timeout: 265 seconds) 2011-10-23T07:57:44 Hi IhateIt, error should be writen in one of log files in directory aichallenge\ants\game_logs\ 2011-10-23T07:57:49 *** bobbydroptable has joined #aichallenge 2011-10-23T07:59:26 there is only 2 files 2011-10-23T07:59:31 one a replay 2011-10-23T07:59:46 and the other o.replay with meaningless content 2011-10-23T08:00:04 it's not in game_logs its the debug.log file in the Tools folder 2011-10-23T08:00:05 *** FusionX has joined #aichallenge 2011-10-23T08:00:06 I opened it in text editor - was meaningless 2011-10-23T08:01:19 IhateIt: what command are you starting it with? if it's "play_one_game", edit it and add the -E parameter to playgame.py call, after that the error will be written to the log files 2011-10-23T08:03:42 IhateIt: its very hard to help "it crashes" 2011-10-23T08:03:45 *** dlila has joined #aichallenge 2011-10-23T08:03:48 I KNOW 2011-10-23T08:03:50 GreenTea: interesting to watch some replays, like http://ash.webfactional.com/replay.624 ... almost parallel ant numbers for quite a while 2011-10-23T08:04:09 *** exezive has joined #aichallenge 2011-10-23T08:04:25 I STATR IT like in tutorial python tools/playgame.py "java -jar ~/AI_CHALLENGE/MyBot.jar" "python tools/sample_bots/python/HunterBot.py" --map_file tools/maps/example/tutorial1.map --verbose -e 2011-10-23T08:04:46 thats from tutrial.sh file 2011-10-23T08:05:07 IhateIt: I think you need the -E written large, Python is case-dependent iirc 2011-10-23T08:05:13 and i dont have any debug.log 2011-10-23T08:05:40 ah - shell i write python with capital as well? 2011-10-23T08:05:52 no, just the -E 2011-10-23T08:06:01 can i write to stdout with my bot without breaking the engine? 2011-10-23T08:06:06 *** ikaros has joined #aichallenge 2011-10-23T08:06:15 *** azEf has joined #aichallenge 2011-10-23T08:06:23 olexs, yes, close game) 2011-10-23T08:06:34 awishformore: stdout is the communication to the server. for debug, better use stderr 2011-10-23T08:07:24 ok 2011-10-23T08:07:31 i tried - it is not this - it still crushes. Is it a way i could present you my source code - it is small and easy. 2011-10-23T08:07:36 though it doesn't answer the question ^^ 2011-10-23T08:08:24 oh my, I might have accidentally created a snake bot http://ash.webfactional.com/replay.625 2011-10-23T08:08:30 no blood: http://ash.webfactional.com/replay.397 2011-10-23T08:10:11 interesting, who is 'A' 2011-10-23T08:11:26 how come that i cannot run anything without crushing it?? It doesn't even start - and I didn't change almost anything from the starter package. Tha's not normal! 2011-10-23T08:13:08 *** avdg has joined #aichallenge 2011-10-23T08:13:15 maybe it's because you hate IT 2011-10-23T08:13:43 This is why I hate it 2011-10-23T08:14:34 oh nice there was a server upgrade?! 2011-10-23T08:15:09 ikaros: there are 3 workers running again, instead of 1 2011-10-23T08:15:10 does anyone use a bot that's not deterministic yet? 2011-10-23T08:15:21 IhateIt: what is the error message. Its python so i doubt its "segfault" what os, what is crashing? your bot or the engine... 2011-10-23T08:15:30 *** avdg1 has quit IRC (Ping timeout: 260 seconds) 2011-10-23T08:15:44 awishformore: my testing ones are not deterministic... ;) 2011-10-23T08:15:48 aka random 2011-10-23T08:16:00 but that does not really count now does it 2011-10-23T08:16:02 yeah well 2011-10-23T08:16:03 delt0r_, python2.5 -c "object.__getattribute__(super, '__bases__')" :-) 2011-10-23T08:16:05 no 2011-10-23T08:16:05 :P 2011-10-23T08:16:06 that's a beginning :) 2011-10-23T08:16:15 i was thinking along the lines 2011-10-23T08:16:26 of nash balancing certain strats 2011-10-23T08:16:29 :> 2011-10-23T08:16:42 bot 0 did not start - thats the message i have 2011-10-23T08:17:00 right and have you used the -E option 2011-10-23T08:17:09 did you get any error file 2011-10-23T08:17:17 and you are using java right? 2011-10-23T08:17:49 in which case what is your java launch command line.. finally how proficient are you with java? 2011-10-23T08:17:49 bot 0 did not start - THIS IS THE MESSAGE THAT I HAVE delt0r - I've no idea why is that? 2011-10-23T08:17:57 *** Areks has joined #aichallenge 2011-10-23T08:17:58 well fine 2011-10-23T08:18:02 good luck 2011-10-23T08:18:03 yes -E 2011-10-23T08:18:10 java right 2011-10-23T08:18:50 2nd year computer science 2011-10-23T08:19:14 i dont have a problems like this ussually 2011-10-23T08:20:06 the tcp server at http://ash.webfactional.com/howto seriously needs more players *hint hint* 2011-10-23T08:21:00 Alexer: what's the deal with the tcp server? 2011-10-23T08:21:08 for good measure, people might run bots simultaneously at http://ants.fluxid.pl/howto and http://ash.webfactional.com/howto 2011-10-23T08:21:42 awishformore, on the official servers you get like.. one game in X hours, right? on tcp servers you get one game in X minutes :P 2011-10-23T08:21:54 so basically testing then 2011-10-23T08:21:56 awishformore, since the load is on the user's computers, not on the server 2011-10-23T08:22:00 awishformore, yep 2011-10-23T08:22:06 ok 2011-10-23T08:22:18 expect my bot to join up in like a week ;) 2011-10-23T08:23:21 but people should leave their bots running so that there's always the critical mass of bots to play good games :P 2011-10-23T08:23:33 *** IhateIt has quit IRC (Quit: Page closed) 2011-10-23T08:23:45 but right now my bot is just sitting around ^^ 2011-10-23T08:24:04 awishformore, my ants just run around like headless chikens :P 2011-10-23T08:24:13 sure :> 2011-10-23T08:24:44 what a coincidence the class i'm working on right now is named "Brain" 2011-10-23T08:25:38 or, actually.. my other ants run around like headless chickens and others behave like horny teenagers: http://ash.webfactional.com/replay.610 2011-10-23T08:26:08 i already caught that one earlier 2011-10-23T08:26:10 does the server have boost installed? I'm rewriting my bot in C++ (from python) and it would be helpful :) 2011-10-23T08:26:38 *** b4Ant has joined #aichallenge 2011-10-23T08:27:03 awishformore, it's a good sign i'm too tired - i've been laughing at it by myself the whole nigh.. i mean morning 2011-10-23T08:27:03 thestinger: Do you rewriting it because of speed? 2011-10-23T08:27:18 yeah 2011-10-23T08:27:22 I'm writing my bot in D by the way :P 2011-10-23T08:28:10 the starter package for D could use some love by the way 2011-10-23T08:29:24 *** parsley_ has joined #aichallenge 2011-10-23T08:29:55 *** boegel has joined #aichallenge 2011-10-23T08:30:51 *** b4Ant has quit IRC (Ping timeout: 256 seconds) 2011-10-23T08:31:47 cirno_the_greate: I'm back, or at least, I'll be reading what you have to say now :) 2011-10-23T08:32:48 Alexer: haha 2011-10-23T08:33:02 oh man, a phallus 2011-10-23T08:33:07 HAHAH 2011-10-23T08:33:08 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-10-23T08:38:17 I use options " -e -E" and write sys.stderr.write("test\n") in bot. File "0.bot0.error" is clean. Where is my "test" string? 2011-10-23T08:38:53 exactly the sa,e problem here 2011-10-23T08:39:25 at least in C++ 2011-10-23T08:39:45 Alexer: ah okay 2011-10-23T08:39:58 Alexer: the spec is not up to date then. it says the server is currently using 55 2011-10-23T08:40:26 there is a preprocessor directive in the Bug.h that comes with the starter package 2011-10-23T08:40:41 #define DEBUG 2011-10-23T08:40:45 which is commented out 2011-10-23T08:40:56 and you need to uncomment to get the debug working 2011-10-23T08:41:28 jmcarthur, yeah, the same information is found on so many places i bet it's hard to remember to update them all :P 2011-10-23T08:41:45 *** nicky40 has quit IRC (Ping timeout: 265 seconds) 2011-10-23T08:42:28 jmcarthur, the local tools were using 55 too until a while ago, apparently. not sure if they got updated, though. 2011-10-23T08:45:09 pedrosorio, for using C++ I need find reason, why pyton tools have no transmit stdin to my C++ program... Debug in C++ works well 2011-10-23T08:46:01 *** danielharan has joined #aichallenge 2011-10-23T08:46:46 i'll be meeting up with mightybyte today. i'll try to convince him to play again 2011-10-23T08:47:09 he had one of the better bots in the tron contest. i don't remember if he participated in planet wars 2011-10-23T08:47:32 *** volkhin has quit IRC (Read error: Connection reset by peer) 2011-10-23T08:47:52 *** danielha_ has joined #aichallenge 2011-10-23T08:48:12 *** delt0r_ has joined #aichallenge 2011-10-23T08:48:19 Belerafon_L: I'm not sure I understand what you mean. Are you using C++? 2011-10-23T08:48:40 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2011-10-23T08:49:00 *** Dispatch has quit IRC (Ping timeout: 265 seconds) 2011-10-23T08:50:02 srsly, why doesnt this work. ive searched through forum and even irc logs. it's vanilla python starter program with "sys.stderr.write("test")" spamming, and -eE switches enabled, creating empty .error files 2011-10-23T08:51:38 *** danielharan has quit IRC (Ping timeout: 258 seconds) 2011-10-23T08:52:13 *** n4nd0 has quit IRC (Ping timeout: 240 seconds) 2011-10-23T08:52:38 pedrosorio: I'm try to use C++ for my bot. Script play_one_game.cmd run without errors but there are no input stream to my.exe. Input function for istream is not called. Now I try to find reason why pyton tools do not worked with my bot. 2011-10-23T08:54:10 Go to the tutorial page, and check the "create test script" section 2011-10-23T08:54:11 http://aichallenge.org/ants_tutorial.php 2011-10-23T08:56:50 *** cowbandit has joined #aichallenge 2011-10-23T08:57:55 *** master_ninja has joined #aichallenge 2011-10-23T08:57:59 hey is there an easy way to see what methods are available to call from ants if you're doing this in java? 2011-10-23T08:58:40 *** ejls has quit IRC (Quit: arg) 2011-10-23T09:00:36 *** flo_ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T09:01:52 *** ejls has joined #aichallenge 2011-10-23T09:05:15 cowbandit: you could always just use an IDE like eclipse or netbeans if you want things like that on demand 2011-10-23T09:05:25 or use a doc generator on the starter package or something 2011-10-23T09:05:36 oo i see 2011-10-23T09:05:54 well i've been using blueJ 2011-10-23T09:06:02 not familiar with it 2011-10-23T09:06:05 so i guess there's no way to see 2011-10-23T09:06:07 ah kk 2011-10-23T09:06:16 (i'm also not really a java developer, so i'm not the best to be giving this advice) 2011-10-23T09:06:46 nice thing about the ide approach is that it's always up to date with your code :D 2011-10-23T09:06:48 nah lol, any advice is appreciated 2011-10-23T09:06:48 *** helpNeeded has joined #aichallenge 2011-10-23T09:07:11 ok i'll check it out 2011-10-23T09:07:49 Hello !!! I just downloaded starter package in java and tried to run it like in tutorial - it didn't even start - it crushed 2011-10-23T09:08:03 "bot 0 did not start" 2011-10-23T09:08:24 man lots of people having problems with starter packages. i wish i was more familiar with them 2011-10-23T09:08:24 "status crashed 0 survived" 2011-10-23T09:08:51 So is there a "cure" ? 2011-10-23T09:09:52 helpNeeded: Try to edit your play_one_game.cmd (.sh) to launch only copies of your bot. 2011-10-23T09:10:55 And put your bot right in the same directory where your play_one_game is. 2011-10-23T09:11:19 Does that game launches? 2011-10-23T09:11:43 helpNeeded: Are you under Windows? 2011-10-23T09:12:24 im testing now 2011-10-23T09:12:28 no - linux 2011-10-23T09:12:55 it doesnt 2011-10-23T09:13:00 helpNeeded: My idea is that %~dp0 does not work with subfolders, i.e. when you write %~dp0\sample_bots\bot.exe it doesn't work. 2011-10-23T09:13:18 *** olexs1 has joined #aichallenge 2011-10-23T09:13:44 %~dp0\sample_bots\bot.exe - there is no such a line anywhere?? 2011-10-23T09:13:48 not on linux 2011-10-23T09:13:51 *** Belerafon_L has quit IRC (Ping timeout: 248 seconds) 2011-10-23T09:14:01 *** he_the_great has joined #aichallenge 2011-10-23T09:14:02 *** boegel has joined #aichallenge 2011-10-23T09:14:08 *** olexs has quit IRC (Ping timeout: 260 seconds) 2011-10-23T09:16:39 *** n4nd0 has joined #aichallenge 2011-10-23T09:18:00 bah, I just registered, went trough tutorial and implemented simple path finding algorithm... in python... to slow, bot timeouts 2011-10-23T09:18:12 *** dabino has joined #aichallenge 2011-10-23T09:18:14 *** Belerafon_L has joined #aichallenge 2011-10-23T09:18:27 is it really possible to do good path finding in python or should i just use C++? Or does my algorithm suck? 2011-10-23T09:18:31 master_ninja: i'm skeptical that python is the reason. 2011-10-23T09:18:43 master_ninja: it's really possible to do a good, fast pathfinding algorithm in any language 2011-10-23T09:19:17 *** lavalamp has joined #aichallenge 2011-10-23T09:19:23 i'm not saying your algorithm sucks, but maybe you should add some caching? 2011-10-23T09:19:33 ah, so brute force algorithm is not a good choice? better check wikipedia for path finding then... 2011-10-23T09:19:37 assuming you have the ram to spare 2011-10-23T09:19:47 master_ninja: brute force never good 2011-10-23T09:19:49 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T09:19:53 cirno_the_greate: ping! 2011-10-23T09:19:57 i'm not even sure what you mean by "brute force" for path finding 2011-10-23T09:20:11 master_ninja: in the same case as your python implementation timeouts on 10 ants, c++ analogue will timeout on 30 ants 2011-10-23T09:20:14 boegel: pong! 2011-10-23T09:20:29 well basically I construct a tree from start and expand to every direction until i reach destination 2011-10-23T09:20:30 is there any way to make the starter package bot working?? 2011-10-23T09:20:35 boegel: which helper functions you need? 2011-10-23T09:20:43 cirno_the_greate: that's assuming his algorithm is O(n). it's almost certainly more expensive than that 2011-10-23T09:20:49 otherwise it wouldn't time out 2011-10-23T09:20:53 cirno_the_greate: I was just going to implement all the helper functions listed at https://github.com/aichallenge/aichallenge/wiki/Ants-Starter-Pack-Guide 2011-10-23T09:21:06 *** kronos has quit IRC (Quit: leaving) 2011-10-23T09:21:15 master_ninja: In the last contest some python bots got a very good places so I think it's quite possible to make heavy calculations with it. 2011-10-23T09:21:23 helpNeeded: how is it not working for you? also, which language? 2011-10-23T09:21:45 cirno_the_greate: we can agree on who implements which functions 2011-10-23T09:21:48 java - i try to test a starter bot 2011-10-23T09:21:53 boegel: passable and unoccupied needs ants state, no? 2011-10-23T09:22:05 status crashed crashed 2011-10-23T09:22:12 thats all i get 2011-10-23T09:22:30 master_ninja: Are you sure that you have wrote a BFS, not a search into the deep :) 2011-10-23T09:22:39 there is only direction missing 2011-10-23T09:22:48 and passable/unoccupied 2011-10-23T09:23:05 and may be visible 2011-10-23T09:23:31 Maybe you are adding new nodes to the beginning of the nodes list while adding to the end is the right implementation. 2011-10-23T09:23:33 UncleVasya: what is BFS? sorry, i really should learn more about path finding it seems 2011-10-23T09:24:07 cirno_the_greate: passable is already there, implemeting occupied and unoccupied is almost trivial 2011-10-23T09:24:09 weird bug 2011-10-23T09:24:11 master_ninja: breadth-first-search, the name of the described by you alhorithm. 2011-10-23T09:24:15 my ants all move on every other turn 2011-10-23T09:24:16 cirno_the_greate: in fact, I just wrote those two functions 2011-10-23T09:24:35 Hi all. Guys, for those interested in pathfinding, you can visit ai-class.org, a well timed Stanford course on AI, and they start exactly with BFS and A* algorithms 2011-10-23T09:24:50 UncleVasya: yeah, just googled it, and it seems that's something I made 2011-10-23T09:24:56 !paste 2011-10-23T09:25:00 sigh i gave up on doing the hoemwork for this week :D 2011-10-23T09:25:10 is there a paste bin? 2011-10-23T09:25:24 boegel: oh, i read about passable 2011-10-23T09:25:27 bobbydroptable: why? it's not *that* hard imho, and you have time until Tuesday 2011-10-23T09:25:36 boegel: use hpaste.org 2011-10-23T09:25:57 bobbydroptable: nooooooooooooooooooooooooooooooooooooo go check the videos on khanacademy (yeah, they don't really teach probability in the aiclass) 2011-10-23T09:26:13 my starter bot doesnt run - i will appreciate any help! "status crashed crashed" this is the message i receive. I am using java 2011-10-23T09:26:14 khanacademy is awesome, just saying 2011-10-23T09:26:20 cirno_the_greate: see http://hpaste.org/53015 2011-10-23T09:26:22 Has anyone submitted java source code? 2011-10-23T09:26:29 cirno_the_greate: the "passable" was already there 2011-10-23T09:26:39 *** danielha_ has quit IRC (Remote host closed the connection) 2011-10-23T09:26:44 well, here's my path finding code http://pastebin.com/fgdr0uU8 if some of you experts instantly see what's slow about it, i'd be glad to hear it :) 2011-10-23T09:26:51 *** avers_ has joined #aichallenge 2011-10-23T09:27:04 boegel: nope, this passable just checks for water 2011-10-23T09:27:18 *** danielharan has joined #aichallenge 2011-10-23T09:27:20 it doesn't mean that it is visible tile, and we can construct path to it 2011-10-23T09:27:28 cirno_the_greate: which is the definition of passable? 2011-10-23T09:27:39 passable means not WATER and ants might have the ability to move there at some point, if not this turn. 2011-10-23T09:28:08 cirno_the_greate: so, this definition is incomplete? 2011-10-23T09:28:38 cirno_the_greate: I don't see how the current implementation doesn't match the description 2011-10-23T09:28:49 *** wartalker has joined #aichallenge 2011-10-23T09:28:53 boegel: may be i'm overthinking 2011-10-23T09:28:54 cirno_the_greate: imho, all you need to do is check whether the tile is water or not 2011-10-23T09:29:03 cirno_the_greate: these are just very basic functions 2011-10-23T09:29:13 cirno_the_greate: also, you don't *have" to use them ;) 2011-10-23T09:29:14 boegel: yep 2011-10-23T09:29:20 i know :) 2011-10-23T09:29:27 also direction function 2011-10-23T09:29:34 cirno_the_greate: let me compile a list of stuff that's missing 2011-10-23T09:29:47 boegel: that would be great 2011-10-23T09:29:57 cirno_the_greate: direction is there, right? as part of the Order datatype? 2011-10-23T09:30:07 no, we need direction function 2011-10-23T09:30:17 Direction: Provide a function to determine the 1 or 2 orthogonal directions from one point to another on a torus. 2011-10-23T09:31:23 boegel: this is from my working code - http://hpaste.org/53016 2011-10-23T09:31:25 pure mess 2011-10-23T09:31:42 cuz doing bot was more fun 2011-10-23T09:31:42 lol 2011-10-23T09:32:47 wow all the ants died 2011-10-23T09:32:49 ==" 2011-10-23T09:32:53 cirno_the_greate: brb 2011-10-23T09:32:58 people's bots are getting better 2011-10-23T09:34:18 *** master_ninja has quit IRC (Quit: gotta go :/ ---) 2011-10-23T09:34:26 *** parsley_ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T09:34:26 Momo beat me :'( 2011-10-23T09:35:53 *** Lore_ has joined #aichallenge 2011-10-23T09:35:56 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T09:36:04 the organization names of the top 6-7 make an hourglass shape right now 2011-10-23T09:36:28 *** Lore_ has quit IRC (Client Quit) 2011-10-23T09:37:02 *** lostUnixForrest has joined #aichallenge 2011-10-23T09:37:35 greetings 2011-10-23T09:38:08 lostUnixForrest, i just checked, i still have tree 2011-10-23T09:38:37 *** moises has joined #aichallenge 2011-10-23T09:38:47 *** lostUnixForrest has left #aichallenge 2011-10-23T09:39:22 by the way. look http://ai-contest.com/visualizer.php?game=3557&user=398 maze is not symmetric 2011-10-23T09:39:24 *** Kingpin13 has joined #aichallenge 2011-10-23T09:39:34 *** wartalker has quit IRC (Quit: Page closed) 2011-10-23T09:40:03 cirno_the_greate: with my implementation of occupied and unoccupied, we only need the 'direction' and 'visible' functions 2011-10-23T09:40:15 GreenTea, rotate 90 and flip 2011-10-23T09:40:32 cirno_the_greate: I'll have to run in 10m, so if you have time to look into direction and visible, let me know 2011-10-23T09:40:48 *** n4nd0 has quit IRC (Ping timeout: 245 seconds) 2011-10-23T09:40:56 i meen, food 2011-10-23T09:41:04 Momo see 2 pices of food 2011-10-23T09:41:14 on fist turn 2011-10-23T09:41:17 and im 1 2011-10-23T09:41:50 ah, interesting 2011-10-23T09:41:56 oh sory 2011-10-23T09:42:09 also 2 %) 2011-10-23T09:42:35 *** avers_ has joined #aichallenge 2011-10-23T09:42:42 *** Fandekasp has joined #aichallenge 2011-10-23T09:42:48 hi 2011-10-23T09:42:54 new member 2011-10-23T09:42:55 yea, i wasnt sure if that one was shaded or not :P 2011-10-23T09:43:17 oopse, good exlpanation for the fail is not correct :D 2011-10-23T09:43:26 hello GREEN TEA - I THINK YOU ARE USING JAVA - right? 2011-10-23T09:43:39 *** nishant has joined #aichallenge 2011-10-23T09:43:42 boegel: i replied to you at github 2011-10-23T09:43:47 GreenTea: it's symmetric. it's mirror symmetry and translated 2011-10-23T09:43:48 i'm going away too 2011-10-23T09:43:50 *** nishant has left #aichallenge 2011-10-23T09:44:00 *** nishant has joined #aichallenge 2011-10-23T09:44:05 could write code for direction at night 2011-10-23T09:44:08 err, it might not be mirror symmetry 2011-10-23T09:44:11 cirno_the_greate: ok, I'll just fork your repo, and issue a pull request 2011-10-23T09:44:11 or may be tomorrow morning 2011-10-23T09:44:12 kind of hard to tell, actually 2011-10-23T09:44:15 helpNeeded, yes 2011-10-23T09:44:18 but i'm certain that it is symmetric 2011-10-23T09:44:21 thanks 2011-10-23T09:44:21 boegel: ok 2011-10-23T09:44:30 how do u make it run on linux?? 2011-10-23T09:44:31 im getting an error when ry to ru my tutorial.sh scrpt 2011-10-23T09:44:38 starter package doesnt run 2011-10-23T09:44:41 guys, at http://aichallenge.org/ants_tutorial_step_5.php part "The Code" there is a missing part to the text: "Put the following code after the food gathering section and before the" . Before the what ? I can't put the code to get the same result as in the replay 2011-10-23T09:44:42 ah, it mirror symmetry 2011-10-23T09:44:43 whoever finishes first can then issue a pull request to qacek 2011-10-23T09:44:47 it says something about deleting a file called singleton lock 2011-10-23T09:44:53 *is 2011-10-23T09:45:05 anyone? 2011-10-23T09:45:12 hm 2011-10-23T09:45:23 how can i debug with the python test suite if i have my executable? 2011-10-23T09:45:26 HOW DO U MAKE IT RUN ON LINUX?? STARTER PACKAGE DOESNT RUN 2011-10-23T09:45:28 where do my exec streams go? 2011-10-23T09:45:30 im just trying to run the starter files for now 2011-10-23T09:45:34 helpNeeded: stop doing all caps like that 2011-10-23T09:45:41 ok 2011-10-23T09:45:46 helpNeeded, im on windows 2011-10-23T09:45:50 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T09:45:54 starter package works for me on linux 2011-10-23T09:46:03 too bad thanks 2011-10-23T09:46:14 starter package asks for a script called tutorial.sh 2011-10-23T09:46:25 m not able to figure otu its contents 2011-10-23T09:46:34 anyone 2011-10-23T09:46:50 *** choas has quit IRC (Ping timeout: 258 seconds) 2011-10-23T09:47:15 How do I get my executable streams in the python test suite? 2011-10-23T09:47:37 *** helpNeeded has quit IRC (Quit: Page closed) 2011-10-23T09:47:42 *** Andsoon has joined #aichallenge 2011-10-23T09:47:57 nishant: http://aichallenge.org/ants_tutorial.php in the "create test script" section 2011-10-23T09:47:59 ah never mind 2011-10-23T09:48:06 it seems to just crash before getting to the first output ;) 2011-10-23T09:49:30 *** danielha_ has joined #aichallenge 2011-10-23T09:49:46 Hi everyone =) 2011-10-23T09:49:50 hey does anyone know how the website's compiler compiles the bots? 2011-10-23T09:49:58 because i'm able to compile my bot on my own computer 2011-10-23T09:50:03 however the website can't seem to compile it 2011-10-23T09:51:17 *** bobbydroptable has quit IRC (Quit: Page closed) 2011-10-23T09:51:21 On a related topic, does anyone know the version of gcc used? 2011-10-23T09:51:34 *** Andsoon has quit IRC (Client Quit) 2011-10-23T09:52:23 *** boegel has quit IRC (Quit: This computer has gone to sleep) 2011-10-23T09:52:35 *** danielharan has quit IRC (Ping timeout: 258 seconds) 2011-10-23T09:52:44 ejls: it says here http://ai-contest.com/starter_packages.php 2011-10-23T09:53:07 *** avers_ has joined #aichallenge 2011-10-23T09:54:03 cowbandit: have the same problem with my java code 2011-10-23T09:54:51 thestinger: oh, thank, I didn't noticed it. 2011-10-23T09:55:44 another issue with java starter package is that it does not recognize hills as hills if there is an ant standing over them. 2011-10-23T09:56:00 *** mceier has joined #aichallenge 2011-10-23T09:56:30 *** cybsy has joined #aichallenge 2011-10-23T09:56:49 *** jandor has joined #aichallenge 2011-10-23T09:56:57 hello people 2011-10-23T09:57:01 hey guys, what's the maximum rows and columns that a map can be? 2011-10-23T09:57:07 are there anybody from russia? 2011-10-23T09:57:40 200x200 if I remember right 2011-10-23T09:57:47 Six: thanks 2011-10-23T09:57:48 read the specification the other day 2011-10-23T09:58:38 *** nishant has quit IRC (Quit: Ex-Chat) 2011-10-23T09:59:24 jandor: aga 2011-10-23T10:00:01 ugh, keep getting failed tests on upload that the test script isn't catching client-side 2011-10-23T10:00:19 wonder if binaries/zips are getting swoggled around somewhere 2011-10-23T10:00:35 (on my end) 2011-10-23T10:02:02 *** joapuipe has joined #aichallenge 2011-10-23T10:02:06 hi 2011-10-23T10:02:12 *** vDevXY has joined #aichallenge 2011-10-23T10:02:40 I submitted my code, but I got this error: "ImportError: No module named ants" 2011-10-23T10:02:46 I'm using Python, btw 2011-10-23T10:03:03 joapuipe: did you include ants.py? 2011-10-23T10:03:09 in zip package? 2011-10-23T10:03:13 aham... 2011-10-23T10:03:18 Nops :) 2011-10-23T10:03:21 :) 2011-10-23T10:03:26 ok, 2011-10-23T10:03:26 *** otzi has joined #aichallenge 2011-10-23T10:03:32 dabino, Ahh - I wondered about that... curious, at least you can change the starter package 2011-10-23T10:03:36 so do I need to include anything else? 2011-10-23T10:03:55 Hmm, the output of the failed test cases is truncated out of usefulness for me 2011-10-23T10:03:58 dabino, I'm rewriting the haskell starter package and it makes the same assumption 2011-10-23T10:04:07 I'm using C# and I would like to use the Location as a struct (for my A* algorithm I need value types) 2011-10-23T10:04:15 did anyone else did this? 2011-10-23T10:04:21 *do 2011-10-23T10:04:24 oh, finally got two games in aichallenge 2011-10-23T10:04:42 Fluxid: do I have to include the "tools" directory or just MyBot.py and ants.py? 2011-10-23T10:04:44 and got rank 30... and it supposedly fell down, so i was higher in rank? O_o 2011-10-23T10:04:51 joapuipe: only code you use 2011-10-23T10:04:56 ok 2011-10-23T10:04:58 thanks 2011-10-23T10:05:13 if yoy MyBot.py uses only module ants, then include ants.py 2011-10-23T10:05:44 ok, thank you! 2011-10-23T10:06:16 Saulzar: that's the problem: game specification is odd. According to it lower letter are bots, capitals are bots on hills. My starter package just recognize all capital Hs as hills... Obviously incorrect, but bots are fighting somehow. 2011-10-23T10:06:46 *** tr3v has joined #aichallenge 2011-10-23T10:07:12 *** tr3v is now known as Guest3544 2011-10-23T10:07:43 maybe my tools are out of date 2011-10-23T10:08:13 *** pitchforkHAULER has joined #aichallenge 2011-10-23T10:08:52 So, first we need clear game specification in plain English, then we can fix starter packages somehow. 2011-10-23T10:09:04 http://aichallenge.org/visualizer.php?game=3579&user=757 i;m surprised how ants fails to raze sir_macelon's hill at (61,52) most the time 2011-10-23T10:09:48 two enemy ants almost razed it but KILLED EACH OTHER right over the hill! 2011-10-23T10:10:34 and turn before ant would raze it, suddenly ant spawns from it to immediately kill attacker 2011-10-23T10:10:37 pure luck 2011-10-23T10:10:43 *** jtamer has quit IRC (Ping timeout: 260 seconds) 2011-10-23T10:11:22 are there any games where someone loses their hill but still wins? 2011-10-23T10:11:43 Fluxid: this is perfect def startegy btw: this is the only way to prevent hill raze vs large army: to spawn ants right in time 2011-10-23T10:12:32 dabino: yeah, but you have no control over spawning 2011-10-23T10:12:46 *** sigh has quit IRC (Remote host closed the connection) 2011-10-23T10:12:55 and because hill is not in view range of any of his ants, then he doesn't know enemy is coming close to hill 2011-10-23T10:13:10 so it's pure coincidence/luck, not strategy 2011-10-23T10:13:23 Fluxid, you have some control over spawning 2011-10-23T10:13:31 you could guard th food until you want to spawn :P 2011-10-23T10:13:35 ok, you can get food on the right time 2011-10-23T10:13:38 noone using c# here? ;) 2011-10-23T10:13:54 but still, you don't know anything about area around the hill 2011-10-23T10:14:56 and i think noone would be so crazy to delay ant spawning just to do "aha, surprise! GOTCHA!" 2011-10-23T10:16:12 *** jandor has quit IRC (Quit: Page closed) 2011-10-23T10:16:21 Fluxid: who knows. that is most interesting part of the show, when all thee little tricks start to appear 2011-10-23T10:17:13 is map format disxription correct there http://aichallenge.org/specification.php ??? 2011-10-23T10:17:53 *** goffrie has joined #aichallenge 2011-10-23T10:18:04 Fluxid: You can share interesting moments this way: http://aichallenge.org/visualizer.php?game=3579&user=757&turn=98&row=61&col=52 2011-10-23T10:18:23 oh, cool! 2011-10-23T10:18:45 *** mrcrusher has joined #aichallenge 2011-10-23T10:19:37 *** twymer has joined #aichallenge 2011-10-23T10:20:10 *** Savaron has quit IRC (Quit: Page closed) 2011-10-23T10:20:27 Anyone knows what is "primary" language, in which starter package all the changes are already implemented 100% correctly? I think I need some back engineering here... 2011-10-23T10:20:32 *** Savaron has joined #aichallenge 2011-10-23T10:20:40 wow, two enemy ants collided on top of the hill? 2011-10-23T10:21:05 the python one seems like the best based on the 3-4 I've tried 2011-10-23T10:21:12 *** Guest3544 has quit IRC (Ping timeout: 240 seconds) 2011-10-23T10:21:24 I think the game was written in python 2011-10-23T10:21:32 what changes 2011-10-23T10:21:32 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T10:21:35 *** joapuipe has quit IRC (Quit: Page closed) 2011-10-23T10:21:48 Six, no, they killed each other. they weren't the same color 2011-10-23T10:22:26 ah, and one died just as it moved onto the hill 2011-10-23T10:22:27 awishformore: I don't know what changes, but java starter package does not read game output correctly 2011-10-23T10:22:38 just before 2011-10-23T10:22:43 som I need to understand the right wat 2011-10-23T10:22:50 *** Laie has joined #aichallenge 2011-10-23T10:23:35 *** n4nd0 has joined #aichallenge 2011-10-23T10:23:50 cyphase: looks like it had a move order to go on the hill and it died during that turn 2011-10-23T10:23:57 does move phase come before attack? 2011-10-23T10:24:13 yes 2011-10-23T10:26:48 GreenTea: do you use official java starter package? 2011-10-23T10:26:59 *** Six has quit IRC (Quit: ☆ï¼äººâ—• ‿‿ ◕人＼) 2011-10-23T10:27:01 combat happens automatically? 2011-10-23T10:27:43 hm 2011-10-23T10:28:03 dabino, take old oficial package in early beta, and then rewrite it 2011-10-23T10:28:39 *** avers_ has joined #aichallenge 2011-10-23T10:28:45 Synchronous swimming: http://aichallenge.org/visualizer.php?game=3695 2011-10-23T10:29:32 yay for unmodified starters! 2011-10-23T10:29:55 GreenTea: thanks for clarification. 2011-10-23T10:29:57 is there some way for the ant to detect that it's hitting a wall? 2011-10-23T10:30:17 isn't there only water? 2011-10-23T10:30:44 <_flag> Fluxid: Can you update the maps on your server? 2011-10-23T10:30:52 or water lol 2011-10-23T10:31:10 *** ants9238483 has joined #aichallenge 2011-10-23T10:31:11 _flag: pull from git or should i put something by myself? 2011-10-23T10:31:11 *** soar has joined #aichallenge 2011-10-23T10:31:43 <_flag> My biggest problem is with games like this: http://ants.fluxid.pl/replay.2141 2011-10-23T10:31:51 *** Dispatch has joined #aichallenge 2011-10-23T10:31:51 *** helkafen has quit IRC (Quit: Page closed) 2011-10-23T10:31:53 cowbandit: there should be a GetIsPassable() in your starter packet 2011-10-23T10:32:02 cowbandit: depending on the language 2011-10-23T10:32:08 ohh ok thanks 2011-10-23T10:32:11 <_flag> The game ends because "food is not being gathered" but really the map is just ridiculously long and they've been changed on the main server so this doesn't happen 2011-10-23T10:32:18 anyone know a way to have your code to send debugging info back from the server? such as time per turn? 2011-10-23T10:32:27 <_flag> So basically I'd just like the maps from the main server 2011-10-23T10:32:30 so is that the recommended way for getting ants unstuck? 2011-10-23T10:34:14 Savaron: what about just "isWater" which i belive it does have at least in some stater packs 2011-10-23T10:34:14 _flag: okay, lemme check which are official on main server 2011-10-23T10:34:21 any c++ devs know if boost is available? 2011-10-23T10:34:41 *** wose has joined #aichallenge 2011-10-23T10:35:20 yea the food not gathered rule needs revising... 2011-10-23T10:35:58 i'm killing tcp server for a moment, sorry 2011-10-23T10:36:04 *** choas has joined #aichallenge 2011-10-23T10:40:16 will it sway competition results if I enter a poor bot in the beginning? I have time next weekend to work on it.. but just not this week 2011-10-23T10:41:20 yay 2011-10-23T10:41:24 my bot is collecting stuff 2011-10-23T10:41:24 ok, server restarted 2011-10-23T10:41:35 now maps are same as on official server 2011-10-23T10:41:48 now there's "only" the combat part left ;) 2011-10-23T10:42:31 @ants, you can log on your local machine if you use the tcp servers to test, since it's running on your local machine 2011-10-23T10:42:32 Dispatch: I have no idea what you mean. 2011-10-23T10:42:39 *** jtamer has joined #aichallenge 2011-10-23T10:42:47 what is the contestbot??? 2011-10-23T10:42:56 *** jtamer_ has joined #aichallenge 2011-10-23T10:43:02 Dispatch: a bot ;) 2011-10-23T10:43:18 hah, I am learning I am not to good with bots... 2011-10-23T10:43:33 *** dom7b5 has quit IRC (Ping timeout: 265 seconds) 2011-10-23T10:43:46 Dispatch: it's an ant bot intelligent enough to talk with people on irc 2011-10-23T10:44:13 lol, except it doesn't know what I mean 2011-10-23T10:44:15 like eliza 2011-10-23T10:44:38 contestbot: hello 2011-10-23T10:44:39 Fluxid: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2011-10-23T10:44:49 Yep, you know we are all here love to write bots. So it's obvious to have our own chatbot here. 2011-10-23T10:45:06 Fluxid: isn't food spawn rate a little bit slow on your server? 2011-10-23T10:45:31 cirno_the_greate: no idea, i just pulled it from git and ran 2011-10-23T10:45:31 contestbot: show us how you can scare stupid people. 2011-10-23T10:45:32 UncleVasya: User error, it's not my fault. 2011-10-23T10:45:34 also some maps which designed for 4 and 8 people runs with 2. As it looks for me. Many mazes, for examples 2011-10-23T10:45:40 Fluxid: ah, okay 2011-10-23T10:45:54 cirno_the_greate: also, on my server view range is larger i think 2011-10-23T10:46:05 lemme check 2011-10-23T10:46:15 fail, I was hoping he answer his 'run away' response :) 2011-10-23T10:46:59 *** jtamer has quit IRC (Ping timeout: 260 seconds) 2011-10-23T10:50:00 soar, You can find out by uploading a small test .. 2011-10-23T10:50:02 *** Chris_0076 has quit IRC (Read error: Connection timed out) 2011-10-23T10:50:13 cirno_the_greate: you're right, settings were different 2011-10-23T10:50:16 if it compiles well I guess it does 2011-10-23T10:50:36 sorry, restarted server again 2011-10-23T10:50:48 *** soar has quit IRC (Quit: Page closed) 2011-10-23T10:50:55 *** Chris_0076 has joined #aichallenge 2011-10-23T10:51:11 i've set similar values as defaults in playgame.py 2011-10-23T10:52:22 Fluxid, http://ai-contest.com/game_settings.php 2011-10-23T10:53:07 g0llum: oh shit... 2011-10-23T10:54:20 but you need more loattime/turntime for yours, of course 2011-10-23T10:57:07 g0llum: only view range ant turns are different 2011-10-23T10:57:15 btw, what is loadtime? 2011-10-23T10:57:39 pedrosorio: time in which bot must load itself before turns begin 2011-10-23T10:58:49 turn time 500 ooooOO 2011-10-23T10:58:55 *** Cyndre has quit IRC (Ping timeout: 248 seconds) 2011-10-23T10:59:04 GreenTea: too little? 2011-10-23T10:59:05 GreenTea: yeah, i was surprised too 2011-10-23T10:59:10 yeah)) 2011-10-23T10:59:17 yes 2011-10-23T10:59:20 simulator uses 1s 2011-10-23T10:59:22 I was surprised too. 2011-10-23T10:59:40 I guess it's supposed to make us develop a nice and simple hive mind 2011-10-23T10:59:41 :P 2011-10-23T11:00:00 but you don't have to do everything each turn 2011-10-23T11:00:06 is there a way to check how much time our bots take to process each turn in games ran on the server? 2011-10-23T11:00:21 i don't think so 2011-10-23T11:00:25 *** exezive has quit IRC (Remote host closed the connection) 2011-10-23T11:01:35 *** pitchforkHAULER has quit IRC (Ping timeout: 248 seconds) 2011-10-23T11:06:02 *** cirno_the_greate has quit IRC (Remote host closed the connection) 2011-10-23T11:06:10 *** foobar_ has joined #aichallenge 2011-10-23T11:06:34 cybsy: no. current ranking have no direct influence on final rankings. The submissions are closed *then* the bots duke it out with a reset rank to determine the competition rank 2011-10-23T11:07:15 when will the submissions be closed? 2011-10-23T11:07:23 cybsy: the current ranks is for feedback and development 2011-10-23T11:07:37 dates are on the web site i believe 2011-10-23T11:07:37 those game settings say turns = 1500 but every game is ended at turn 1000? 2011-10-23T11:09:04 *** Fandekasp has quit IRC (Ping timeout: 258 seconds) 2011-10-23T11:10:52 *** Mixailo has joined #aichallenge 2011-10-23T11:10:58 Hi guys 2011-10-23T11:11:02 *** swamp has joined #aichallenge 2011-10-23T11:11:10 hi 2011-10-23T11:11:29 *** deltree_ has joined #aichallenge 2011-10-23T11:11:55 good morning 2011-10-23T11:12:56 *** swamp has left #aichallenge 2011-10-23T11:14:32 thanks delt0r 2011-10-23T11:14:58 *** Blkt has joined #aichallenge 2011-10-23T11:15:14 *** _rbarboza <_rbarboza!bb1636d4@gateway/web/freenode/ip.187.22.54.212> has joined #aichallenge 2011-10-23T11:15:30 a date would be nice 2011-10-23T11:15:55 *** samnon has quit IRC (Quit: Page closed) 2011-10-23T11:17:47 awishformore: are you hitting on me? 2011-10-23T11:18:03 i don't know, pix? 2011-10-23T11:18:07 *** jezuz has joined #aichallenge 2011-10-23T11:18:11 :D 2011-10-23T11:18:23 that's the best i can do in two characters of ascii 2011-10-23T11:18:34 but i assure you, i do have a nose too 2011-10-23T11:19:42 asl? 2011-10-23T11:19:43 *** jezuz` has quit IRC (Ping timeout: 248 seconds) 2011-10-23T11:19:47 *** bmh has joined #aichallenge 2011-10-23T11:19:51 *** bmh has joined #aichallenge 2011-10-23T11:19:56 good morning 2011-10-23T11:20:09 awishformore: i'll let you binary search my asl 2011-10-23T11:20:13 eyes look a bit tiny but the big mouth should make up for that 2011-10-23T11:20:26 i like the way you think :) 2011-10-23T11:21:28 It's a bit of a drag that the Haskell bindings don't support Hills. 2011-10-23T11:22:30 hey what is the method that returns the unexplored squares? 2011-10-23T11:22:57 hmm, been waiting for a game for hours and then my opponent crashed the first round :-( 2011-10-23T11:24:07 :-) 2011-10-23T11:24:13 *** n4nd0 has quit IRC (Read error: Operation timed out) 2011-10-23T11:24:49 love in #aichallenge now we just need the next competition to be love-bots 2011-10-23T11:24:51 I'm mostly ending up with the most ants, at least if the other bots aren't focussed on food (mine doesn't hunt food, it simply goes random) 2011-10-23T11:25:04 I'm actually working on an other bot 2011-10-23T11:25:04 *** Cyndre has joined #aichallenge 2011-10-23T11:27:44 eh 2011-10-23T11:27:48 weird viewradius 2011-10-23T11:27:57 sqrt(77)? 2011-10-23T11:29:08 yeah 2011-10-23T11:29:12 *** onensora has quit IRC () 2011-10-23T11:29:34 and what's all those food vars 2011-10-23T11:29:40 no freakin clue 2011-10-23T11:29:46 depends on api 2011-10-23T11:31:34 hey 2011-10-23T11:31:35 whats up? 2011-10-23T11:32:01 *** yoden has joined #aichallenge 2011-10-23T11:32:05 *** HaraKiri has quit IRC (Ping timeout: 260 seconds) 2011-10-23T11:32:52 *** HaraKiri has joined #aichallenge 2011-10-23T11:34:12 Can you raze your own hill? 2011-10-23T11:34:28 no 2011-10-23T11:34:42 but you can stand on your own hill to prevent spawning 2011-10-23T11:34:55 which you'd probably only want to do if you have more than one hill 2011-10-23T11:36:22 anyone an idea how to beginn as a complete newbie in this area for a c++ bot? :p 2011-10-23T11:36:43 *** _rbarboza <_rbarboza!bb1636d4@gateway/web/freenode/ip.187.22.54.212> has quit IRC (Ping timeout: 265 seconds) 2011-10-23T11:36:45 vDevXY: download the starter package, enter it into the contest? 2011-10-23T11:36:50 i quite dont get the sense of this antsgame 2011-10-23T11:37:24 thought the starter package is too bad and must be modified... but i dont know where to beginn 2011-10-23T11:37:45 vDevXY: the rules are not perfectly clear, but you should probably read them first 2011-10-23T11:37:48 tbh, never liked the starter packages :P 2011-10-23T11:37:59 *** _robz <_robz!bb1636d4@gateway/web/freenode/ip.187.22.54.212> has joined #aichallenge 2011-10-23T11:38:05 prefer the "mega ultra 1 mb source file" approach :P 2011-10-23T11:38:43 vDevXY: the next thing is go to the tutorial page and see how to test your bot in your computer http://aichallenge.org/ants_tutorial.php 2011-10-23T11:39:13 thats the problem, i only know: there are ants, that spread through feed, they kill other ants, must take over hills, but no plan how i get the parameters to test my bot^^ 2011-10-23T11:39:19 thx @pedrosorio 2011-10-23T11:39:40 vDevXY: and then the readme of the starter package which gives a short description of what each source file is doing 2011-10-23T11:42:18 what i also dont understand: we only have 1 second time per turn, dont we? 2011-10-23T11:42:19 Is it cheating to infer the content of the map based on the requirement that maps are symmetric? 2011-10-23T11:43:31 *** pitchforkHAULER has joined #aichallenge 2011-10-23T11:43:41 bmh, i don't think so 2011-10-23T11:43:51 *** moises has quit IRC (Remote host closed the connection) 2011-10-23T11:44:44 hmph, that doesn't make sense 2011-10-23T11:44:50 *** ololo has joined #aichallenge 2011-10-23T11:45:32 is there a method to determine whether a location is in an ant's sight range? 2011-10-23T11:45:41 or do we have to make one 2011-10-23T11:45:47 what do u think is better for a bot ansi c or c++? 2011-10-23T11:45:58 running on linux 2011-10-23T11:46:17 vDevXY: no disrespect meant, but are you a C n00b? 2011-10-23T11:46:42 *** Laie has quit IRC (Ping timeout: 260 seconds) 2011-10-23T11:47:20 never used c on linux, some winapi on windows years ago 2011-10-23T11:47:32 СУИСÐЙД ДЕТОЧКÐ! 2011-10-23T11:47:46 if you don't have a lot of programming experience it would probably be better to take python or so 2011-10-23T11:47:51 I wouldn't use C for this, but then again, I'm using Haskell. 2011-10-23T11:47:57 :) 2011-10-23T11:47:59 (What ikaros said) 2011-10-23T11:48:19 and why haskell? 2011-10-23T11:48:29 *** guest1101 has joined #aichallenge 2011-10-23T11:48:40 @ikaros i didnt programm for about a year but c++ shouldnt be a problem 2011-10-23T11:48:41 vDevXY: User error, it's not my fault. 2011-10-23T11:48:42 *** GeorgeSebastian has joined #aichallenge 2011-10-23T11:49:10 <_robz> hi...is there any way to reset the passwd. i've just activated my account and although i'm pretty sure username/passwd are correct, this is probably the easiest solution... 2011-10-23T11:49:16 *** kronos has joined #aichallenge 2011-10-23T11:50:10 how do I change the timeout on the bots when debugging? 2011-10-23T11:50:30 just take the language which you are best at :) 2011-10-23T11:50:41 best hint i can give 2011-10-23T11:50:59 you could always port to a different language later 2011-10-23T11:51:06 yeah c++ shouldnt be a problem else python, but c++ should be better in speed, much better^^ 2011-10-23T11:51:49 for example i'm prototyping with lua and will perhaps change to c++ or ocaml in the end.. not sure.. perhaps lua will be fine for the whole time 2011-10-23T11:51:49 vDevXY: when choosing a language for a project, the most important criteria to me is how fun it will be 2011-10-23T11:52:01 *criterion 2011-10-23T11:52:03 btw.. 2011-10-23T11:52:11 *** ololo has quit IRC (Ping timeout: 265 seconds) 2011-10-23T11:52:15 bmh: haskellers unite! 2011-10-23T11:52:17 *** Racko has quit IRC (Quit: Page closed) 2011-10-23T11:52:25 never heard of lua 2011-10-23T11:52:39 how about submitting zips for compiled languages.. do we have to put a makefile in the zip or how does the server compile stuff 2011-10-23T11:52:54 e.g. if i choose to use ocaml will the server have ocamlbuild? 2011-10-23T11:52:56 ikaros: it looks for a file called MyBot.extension 2011-10-23T11:53:00 yea 2011-10-23T11:53:09 *** guest1101 has quit IRC (Client Quit) 2011-10-23T11:53:10 I thought on taking lisp and learning to program in it while doing the programming. But I backed up :(. I'm doing it in C++ now. 2011-10-23T11:53:19 ikaros: it doesn't really use build systems 2011-10-23T11:53:26 but there has linking to be done.. cant figure that out all alone 2011-10-23T11:54:02 ikaros: https://github.com/aichallenge/aichallenge/blob/epsilon/worker/compiler.py#L302 2011-10-23T11:54:19 nice thx i will look into that 2011-10-23T11:54:21 can we use some extra libraries? like boost? 2011-10-23T11:54:43 that's what i was looking for: "OCaml" : [["ocamlbuild -lib unix", BOT + ".native"]] 2011-10-23T11:54:46 you can embed it into your source 2011-10-23T11:54:46 great :) 2011-10-23T11:54:53 vDevXY: anything the server doesn't have you have to include with your bot. large libraries might be problematic 2011-10-23T11:55:02 use bcp for that so it copies only what you really need from boost 2011-10-23T11:55:30 i love seeing so much functional programming in this contest 2011-10-23T11:56:07 if i'd use c++ i would use headers libs only anyways 2011-10-23T11:56:23 jmcarthur, that's one cool thing about ai contests :) 2011-10-23T11:57:48 *** aquatic has joined #aichallenge 2011-10-23T11:58:14 *** ringsee has joined #aichallenge 2011-10-23T11:59:32 Question about maps: Is every given location accessible from every other given location? Can there be a map split in two where both teams fight on both sides but can't interact? 2011-10-23T12:00:18 *** danielha_ has quit IRC (Remote host closed the connection) 2011-10-23T12:00:57 *** danielharan has joined #aichallenge 2011-10-23T12:00:58 i've seen maps where hills were surrounded by water :) 2011-10-23T12:01:17 ringsee: I don't believe that maps must be connected 2011-10-23T12:01:53 Ok, thanks. 2011-10-23T12:02:50 * avdg believes that 1 bot should be able to win victory 2011-10-23T12:03:00 but thats personal 2011-10-23T12:03:28 s/bot/ant/ 2011-10-23T12:03:44 *** roflmao has joined #aichallenge 2011-10-23T12:03:48 FUCK ME 2011-10-23T12:03:50 I remember reading the maps were supposed to be simply connected 2011-10-23T12:04:07 yeah, me as well 2011-10-23T12:04:20 WHAT THE FUCK MY a* is fine why do you keep on doing 360's you stupid ant 2011-10-23T12:04:22 you stupid stupid stupid ant 2011-10-23T12:05:02 although I wonder if there are some where there are water cells which are never visible (i.e. more than sort(77) from the closest land) 2011-10-23T12:05:38 don't let your bot be fooled by not spotted land ;-) 2011-10-23T12:05:53 how does python compare speedwise with c++? 2011-10-23T12:06:03 could wreak havoc on bots trying to scout all the map 2011-10-23T12:06:20 cause the sample bots in py time out on my comp :P 2011-10-23T12:06:42 ants9238483, what kind of computer? 2011-10-23T12:06:54 its probably more your code behaving like that, if you really want to know ;-) 2011-10-23T12:06:59 2.8 dual core 2011-10-23T12:07:10 *** mrcrusher has quit IRC (Ping timeout: 265 seconds) 2011-10-23T12:07:12 my cpp code runs fine, but the sample bots time out :P 2011-10-23T12:07:20 ants9238483, that's weird.. a python bot is #20, so it shouldn't be that slow 2011-10-23T12:07:22 avdg, true but at 500ms turn time there can be a difference :) 2011-10-23T12:07:41 also true 2011-10-23T12:07:50 turn time is 500ms? 2011-10-23T12:08:00 set it to 300 2011-10-23T12:08:24 8 player multi hill 2011-10-23T12:09:35 otzi if this page is correct: http://aichallenge.org/game_settings.php 2011-10-23T12:09:35 *** aquatic has quit IRC (Read error: Connection reset by peer) 2011-10-23T12:11:50 *** racko has joined #aichallenge 2011-10-23T12:12:16 ants9238483: if you aren't just calling library functions python is at least 20x slower (probably more) but the sample bots don't really do very much so you might have something else going on 2011-10-23T12:12:33 does speed decrease 2011-10-23T12:12:35 with more instances? 2011-10-23T12:12:41 thestinger, do you have psyco installed? 2011-10-23T12:12:44 cause i'm using 7 2011-10-23T12:12:44 more instances of what? 2011-10-23T12:12:49 more instances of bots 2011-10-23T12:13:02 ants9238483 i mean. do you have psyco? 2011-10-23T12:13:32 *** boegel has joined #aichallenge 2011-10-23T12:13:33 ah fuck yes 2011-10-23T12:13:34 idk what that is 2011-10-23T12:13:48 running a mac 2011-10-23T12:13:54 so i know it has python 2011-10-23T12:14:00 but idk if it has that 2011-10-23T12:14:00 what python version is it? 2011-10-23T12:14:02 ants9238483, it's a python module that can greatly speed up execution 2011-10-23T12:14:05 I believe the game should send output to your bots one at a time otherwise it's hard to test 2011-10-23T12:14:25 # turn 1 bot 0 timed out 2011-10-23T12:14:34 does that mean my turn took too long? 2011-10-23T12:15:06 how to check that? 2011-10-23T12:15:48 type python --version 2011-10-23T12:16:00 2.7.1 final 2011-10-23T12:16:11 *** drzhbe has joined #aichallenge 2011-10-23T12:16:16 *** amstan has joined #aichallenge 2011-10-23T12:16:16 *** ChanServ sets mode: +o amstan 2011-10-23T12:16:42 roflmao it can happen if you don't send the "go" for the turns as well. 2011-10-23T12:16:51 ants9238483: does HunterBot.py timeout? 2011-10-23T12:17:07 that's the one 2011-10-23T12:17:16 *** tillj has joined #aichallenge 2011-10-23T12:17:46 i have one instance of MyBot.cc, 4 of hunter, 3 of lefty 2011-10-23T12:17:48 *** dabino has quit IRC (Ping timeout: 265 seconds) 2011-10-23T12:18:20 the go? 2011-10-23T12:19:13 each turn is a sequence of move orders terminated by go on a line by itself. If you don;t send the go for some reason the server can time out waiting for it 2011-10-23T12:19:30 or maybe my A* just takes too long 2011-10-23T12:19:50 hm, that's odd, I'm using python 2.6.6 and it doesn't timeout, even with 300ms turn. Maybe it's the fact that you're running 4 instances. I don't know 2011-10-23T12:20:11 My computer is less powerful than yours 2011-10-23T12:20:14 I don't know my a* easily takes 400ms to calculate the path for my antsy 2011-10-23T12:20:21 well, guess i'll just set turn to liek 2000 ms :P 2011-10-23T12:20:44 roflmao: guess you can work with one ant per turn, then :P 2011-10-23T12:20:52 ojnly have one ant per turn :P 2011-10-23T12:21:11 what's weird is sometimes the warnings appear sometimes they don't 2011-10-23T12:21:49 can I increase thinking time per turn? 2011-10-23T12:22:04 *** Axoren has joined #aichallenge 2011-10-23T12:22:18 you can set --turntime on tools/playgame.py 2011-10-23T12:22:20 anyone know the CLI command for running the TCP client? 2011-10-23T12:22:28 but on server it is fixed right? 2011-10-23T12:22:39 yeah, on server you have no way to set it 2011-10-23T12:22:53 they give you very little time to think then 2011-10-23T12:23:33 *** varikuzhi has joined #aichallenge 2011-10-23T12:25:18 nope you were right 2011-10-23T12:25:19 *** deltree_ has quit IRC (Quit: If you don't like it, don't suck at it) 2011-10-23T12:25:23 wasn't sending go command 2011-10-23T12:26:23 no 2011-10-23T12:26:26 argh 2011-10-23T12:26:28 if you had a situation like this: 2011-10-23T12:26:30 *12* 2011-10-23T12:26:34 2**1 2011-10-23T12:26:38 *12* 2011-10-23T12:26:44 then who would live? 2011-10-23T12:28:31 whatever ants are in range of the least opponents, are the ones who live 2011-10-23T12:29:16 *** tncardoso has joined #aichallenge 2011-10-23T12:29:21 not that that scenario is even really possible 2011-10-23T12:29:23 how can my a* time out with over 5 seconds of time 2011-10-23T12:29:38 for a food like 3 squares away :| 2011-10-23T12:29:44 *** _robz <_robz!bb1636d4@gateway/web/freenode/ip.187.22.54.212> has quit IRC (Quit: Page closed) 2011-10-23T12:30:08 *** Pwnerat0r has joined #aichallenge 2011-10-23T12:31:11 warn @persistent.path.to_s 2011-10-23T12:31:17 seens to always fail if I put taht 2011-10-23T12:31:23 roflmao: I think you search the same squares again and again. 2011-10-23T12:31:30 nope 2011-10-23T12:31:34 I have a closed set 2011-10-23T12:31:50 # Add already checked node to closed set. 2011-10-23T12:31:50 closed_set << node 2011-10-23T12:31:54 # Skip this node if it has already been checked. 2011-10-23T12:31:54 next if closed_set.include? neighbor 2011-10-23T12:32:43 just write all nodes you visit to a file or stderr and check 2011-10-23T12:32:46 *** Redgis has quit IRC (Read error: Connection reset by peer) 2011-10-23T12:32:59 *** drzhbe has quit IRC (Quit: Page closed) 2011-10-23T12:33:13 *** Redgis has joined #aichallenge 2011-10-23T12:33:16 wow 2011-10-23T12:33:19 reconstruct_path(came_from, came_from[goal]) 2011-10-23T12:33:28 that's the line that makes it sometimes time out, the actual a* always runs really quickly as it should 2011-10-23T12:33:31 hilarious 2011-10-23T12:34:12 *** Axoren has quit IRC (Quit: Page closed) 2011-10-23T12:36:07 oh wow 2704 nodes in the opens et 2011-10-23T12:36:10 that's just wrong 2011-10-23T12:36:27 roflmao: roflmao 2011-10-23T12:36:33 xD 2011-10-23T12:36:39 how the fuck did that happen? 2011-10-23T12:38:30 roflmao: Maybe your closed set erases in some place? 2011-10-23T12:39:17 bnah it has 2427 nodes :| 2011-10-23T12:39:47 open_set is prioritized by cost 2011-10-23T12:40:02 *** foobar_ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T12:40:03 cost of heuristic from node + what is already traveled from that node 2011-10-23T12:40:17 so it shouldn't have to go through all 2427 nodes to reach the obvious square in front 2011-10-23T12:40:23 heh, maybe you mixed up < and > ? :D 2011-10-23T12:40:27 maybe :D 2011-10-23T12:40:35 that would be hilarious 2011-10-23T12:41:06 *** ringsee has quit IRC (Quit: Lost terminal) 2011-10-23T12:42:45 *** n4nd0 has joined #aichallenge 2011-10-23T12:43:11 ah, the sorting isn't working 2011-10-23T12:43:51 now just checked 24 something ndoes 2011-10-23T12:43:53 rofl 2011-10-23T12:44:47 roflmao: Once I did that in a model with 240k nodes :) 2011-10-23T12:45:02 (and god knows how many edges) 2011-10-23T12:46:35 *** choas has quit IRC (Ping timeout: 276 seconds) 2011-10-23T12:47:03 all working! 2011-10-23T12:49:01 *** bward has joined #aichallenge 2011-10-23T12:49:30 Is someone using clojure and willing to chat? 2011-10-23T12:53:11 damn, i messed up something in my bot... 2011-10-23T12:53:12 *** Fandekasp has joined #aichallenge 2011-10-23T12:53:18 *** Cyndre_ has joined #aichallenge 2011-10-23T12:53:24 re 2011-10-23T12:53:31 Some python devs here ? 2011-10-23T12:53:49 here 2011-10-23T12:53:50 Fandekasp, what's up? 2011-10-23T12:53:54 *** danielharan has quit IRC (Remote host closed the connection) 2011-10-23T12:54:15 *** Areks has quit IRC (Read error: Connection reset by peer) 2011-10-23T12:54:41 *** Areks has joined #aichallenge 2011-10-23T12:54:48 amstan, hi. Is it hard to implement statistic for a time taken by bot for his turns? I think it would be interesting to people to compare their bots in "place_in_the_ranking_table / average_turn_time" 2011-10-23T12:54:54 In my do_turn(), I'd like to put a breakpoint to study a bit more what's going on, what data are available etc (just finished the tuto). But when I put a ipdb, then run the bot, it gives a "bot 0 timed out". Is there a debug mode or something ? 2011-10-23T12:55:28 *** Cyndre has quit IRC (Ping timeout: 256 seconds) 2011-10-23T12:55:39 UncleVasya: do you know php and python? 2011-10-23T12:55:46 Fandekasp, well any kind of pausing for debugging is gonna time it out, unless you modify the tools 2011-10-23T12:56:00 no, I don't 2011-10-23T12:56:23 well, it'll be mostly python 2011-10-23T12:56:24 cyphase: ok. Didn't you do something like that ? How are you debugging your code ? 2011-10-23T12:56:37 Fandekasp, i haven't started coding yet :) 2011-10-23T12:56:43 How much wall time do we have for each turn? 2011-10-23T12:56:46 lol 2011-10-23T12:57:01 cyphase: are you a aichallenge core dev ? 2011-10-23T12:57:05 although i'm about to get into it, i just had some other stuff to finish 2011-10-23T12:57:08 UncleVasya: yeah, it's hard, i don't think the time per turn is saved anywhere 2011-10-23T12:57:10 Fandekasp, umm, no 2011-10-23T12:57:23 Fandekasp, perhaps you're confusing me with someone :) 2011-10-23T12:57:23 ok 2011-10-23T12:57:24 lol.. how come you answer other people questions then xD 2011-10-23T12:57:33 cyphase: modify run_one_bot, there's a way to give timeout info to the engine 2011-10-23T12:57:37 thanks for the response. 2011-10-23T12:57:54 Fandekasp, i've been poking around and getting a feel for it 2011-10-23T12:58:03 ok 2011-10-23T12:58:25 *** GeorgeSebastian has quit IRC (Ping timeout: 260 seconds) 2011-10-23T12:58:47 Fandekasp: ./playgame.py allows you to specify --turntime and --loadtime 2011-10-23T12:59:19 amstan: oh so a several minutes time will allow me to debug. yes, just logic ^^ 2011-10-23T12:59:23 thx 2011-10-23T12:59:39 so debugmode = --turntime 3600 2011-10-23T13:00:03 *** Cyndre_ has left #aichallenge ("Leaving") 2011-10-23T13:00:30 yep 2011-10-23T13:00:34 hmm no, ipdb doesn't prompt 2011-10-23T13:00:36 *** Cyndre has joined #aichallenge 2011-10-23T13:00:43 I'll have a look at the engine code 2011-10-23T13:01:02 *** deltree_ has joined #aichallenge 2011-10-23T13:01:16 Fandekasp, 3600 would only be 3.6 seconds 2011-10-23T13:01:54 *** Pwnerat0r has quit IRC (Quit: Page closed) 2011-10-23T13:02:42 I see, but I still need to get my prompt ^^ 2011-10-23T13:03:12 *** Larose has joined #aichallenge 2011-10-23T13:04:07 hey all 2011-10-23T13:04:56 *** choas has joined #aichallenge 2011-10-23T13:05:13 Any advice on stopping the bot timing out? I'm running the python tutorial bot and I've added BFS for pathfinding, but it times out quite quickly in any game I play, even if I double the turn time. 2011-10-23T13:06:03 bward: A*? 2011-10-23T13:06:16 bward: try to check only visible locations 2011-10-23T13:07:02 UncleVasya, I've changed the exploring code so it BFSes outwards from all idle ants and goes to the first nonvisible square so I'm not checking everything 2011-10-23T13:07:15 bmh, is it much faster? I tried that but it didn't seem to make a huge difference 2011-10-23T13:07:25 try to finish turn right when all of your ants have been moved. 2011-10-23T13:07:50 Hah, the one time my bot won a game, it timed out. 2011-10-23T13:07:59 *** bmh has quit IRC (Quit: bmh) 2011-10-23T13:08:09 Also, I think the way to go is to search from destinations (food, hill, etc) to ants. 2011-10-23T13:08:20 meh stupid fucking thing is still getting 2880 nodes in the open sector soimetimes 2011-10-23T13:08:27 Yeah that's what I've been doing 2011-10-23T13:08:43 But I can't see a way of doing that for exploring 2011-10-23T13:09:08 I think my food gathering code runs ok, but the exploring really slows it down, even though I've made it quite a bit faster than the tutorial code 2011-10-23T13:09:27 really? 2011-10-23T13:09:31 what do you use for food gathering 2011-10-23T13:09:32 a*? 2011-10-23T13:09:35 *** azEf has quit IRC (Ping timeout: 248 seconds) 2011-10-23T13:09:40 BFS backwards from the food 2011-10-23T13:09:41 bward: I've done it for exploring too. 2011-10-23T13:09:57 UncleVasya, so you search from an unexplored square back to an ant? 2011-10-23T13:10:31 For exploration I make a boundary_list and then launch a search from it to ants. 2011-10-23T13:10:31 No, BFS is ok for me. 2011-10-23T13:11:03 I was thinking of wall hugging ala lefty bot for exploration 2011-10-23T13:11:06 once food has been sighted 2011-10-23T13:11:08 a* 2011-10-23T13:11:13 which leads to a chain reaction for a while 2011-10-23T13:11:16 then back to wall hugging 2011-10-23T13:11:21 Simultaniously from all of them. In one run on the whole map I found all free ants and give them ordes. 2011-10-23T13:11:48 UncleVasya, thanks :) 2011-10-23T13:11:52 Something to think about. 2011-10-23T13:11:57 isn't BFS just an unoptimized a*? 2011-10-23T13:12:05 *** Axoren has joined #aichallenge 2011-10-23T13:12:23 I'm starting to hate the java implementation of the starter files 2011-10-23T13:12:42 I might literally just learn python 2011-10-23T13:13:10 *** peregrinator has joined #aichallenge 2011-10-23T13:13:10 I didn't try to implement A* ever (in fact I implemented BFS for the first time to this contest) :) 2011-10-23T13:13:23 and your bfs doesn't crash after a while? :S 2011-10-23T13:13:35 my thing just grinds to a halt cause I end up with 2000 something nodes on my open set array 2011-10-23T13:13:42 I think that for A* you have to know a start and end point? So it's no use for searching? 2011-10-23T13:13:43 Axoren: OCaml's starter package is the best one :P 2011-10-23T13:13:50 *** AlexZagr has joined #aichallenge 2011-10-23T13:13:55 Hi all. Can anyone point me towards finding out why my bot doesn't pass test_bot.sh ? 2011-10-23T13:13:57 yeah you do 2011-10-23T13:13:58 For java? 2011-10-23T13:14:00 but you always start with food near base 2011-10-23T13:14:03 or python? 2011-10-23T13:14:11 you go to that then if you can see any other food in your LOS you go to that and so on 2011-10-23T13:14:18 if you can't see any other food in LOS go to searching 2011-10-23T13:14:21 that's what I'm planning on doing 2011-10-23T13:14:40 ah, I mean starter package for OCaml language. 2011-10-23T13:14:42 Makes sense. I don't think you could still do that for exploring though? 2011-10-23T13:14:44 OCaml does not sound like a language... 2011-10-23T13:14:46 *** xar0l has joined #aichallenge 2011-10-23T13:14:53 is it a time out error? 2011-10-23T13:14:57 That sounds like the beginning of a horrible pun 2011-10-23T13:15:03 Like "O Caml, you so silly." 2011-10-23T13:15:07 *** pistacchio has quit IRC (Quit: Page closed) 2011-10-23T13:15:41 Is OCaml easy to learn? 2011-10-23T13:15:57 ants9238483: It's SandboxError: Failed to start ['./MyBot.rb'] waiting 0.25 seconds for bots to process end turn 2011-10-23T13:16:07 Let's say it's... interesting :D 2011-10-23T13:16:20 *** lemnar has joined #aichallenge 2011-10-23T13:17:02 but I'm not sure why it's giving that message - it runs fine under play_one_game.sh 2011-10-23T13:17:03 http://www.eatliver.com/img/2010/5926.jpg 2011-10-23T13:17:10 I think I found a correct way to debug 2011-10-23T13:17:19 *** kozlovsky has joined #aichallenge 2011-10-23T13:17:37 using the LogFilter in sample_bots/python/GreedyBot.py 2011-10-23T13:17:43 let me test that 2011-10-23T13:18:02 *** GeorgeSebastian has joined #aichallenge 2011-10-23T13:18:08 I'm testing as well ^^ 2011-10-23T13:18:42 *** AlexZagr has quit IRC (Ping timeout: 265 seconds) 2011-10-23T13:19:11 it's funny my a* i created standalone doesn't explode with nodes 2011-10-23T13:19:13 but this one does 2011-10-23T13:19:31 think i know what it is 2011-10-23T13:19:41 instead of './MyBot.rb' 2011-10-23T13:19:43 For logging I output to a file from my bot (which I turn off with a flag before submitting) 2011-10-23T13:19:51 just do 'ruby MyBot.rb' 2011-10-23T13:20:22 so ants the command I ran was: ./test_bot.sh ../ruby_ant/MyBot.rb 2011-10-23T13:20:47 *** AlliedEnvy has quit IRC (Ping timeout: 248 seconds) 2011-10-23T13:20:48 should I move my bot into the tools directory? 2011-10-23T13:20:59 that would make life simpler, yes 2011-10-23T13:21:05 if you do that 2011-10-23T13:21:09 your command should be 2011-10-23T13:21:17 ./test_bot.sh 'ruby MyBot.rb' 2011-10-23T13:21:26 ah I see! 2011-10-23T13:22:43 *** deltree_ has quit IRC (Quit: If you don't like it, don't suck at it) 2011-10-23T13:22:48 ok so now I get similar output as to the game when I submit - but I'm not sure I see errors 2011-10-23T13:22:49 *** lemnar has quit IRC (Quit: Page closed) 2011-10-23T13:23:02 btw, windows guys can use this for comfort testing: http://www.filefactory.com/file/ce6905e/n/AntsTesting_0_5.7z Let me know if you have problems with it. 2011-10-23T13:23:25 can you link me the engine output? 2011-10-23T13:23:33 UncleVasya, cool, will take a look :) 2011-10-23T13:24:13 ants: https://gist.github.com/1307602 2011-10-23T13:24:46 it seems 2011-10-23T13:24:53 that you're echoing all imput you get 2011-10-23T13:25:18 if you want to do that, best to do it in stderr or in a debug file 2011-10-23T13:25:45 oh I see - that was in the example ruby bot I started with 2011-10-23T13:26:09 I didn't bother removing that - assumed it was ok to be there 2011-10-23T13:26:10 The time of uploading has came. 2011-10-23T13:26:13 *** bitbrit has quit IRC (Quit: bitbrit) 2011-10-23T13:26:19 it passes now! Thanks ant! 2011-10-23T13:26:39 *** racko has quit IRC (Ping timeout: 248 seconds) 2011-10-23T13:26:41 *for ? 2011-10-23T13:27:10 no prob 2011-10-23T13:27:30 *** racko has joined #aichallenge 2011-10-23T13:27:47 *** SqrT__ has joined #aichallenge 2011-10-23T13:29:47 *** azEf has joined #aichallenge 2011-10-23T13:29:59 *** bitbrit has joined #aichallenge 2011-10-23T13:30:05 fixed my a8 2011-10-23T13:30:22 *** dr- has quit IRC (Quit: Do you like fishsticks ? You are a gay fish !) 2011-10-23T13:30:40 What was the problem? 2011-10-23T13:31:44 I was using the wrong algorithm for the priority queue sorting 2011-10-23T13:33:30 still spikes occasionally ned to figure it out haha 2011-10-23T13:34:13 it's fricken insanely great when it works though :P 2011-10-23T13:34:21 15 ants by turn 100 2011-10-23T13:35:26 *** vDevXY has quit IRC (Quit: Page closed) 2011-10-23T13:35:34 it's sad to come to the last page from the first one... 2011-10-23T13:41:07 *** AlliedEnvy has joined #aichallenge 2011-10-23T13:42:54 *** than_ has joined #aichallenge 2011-10-23T13:43:18 *** n4nd0 has quit IRC (Ping timeout: 244 seconds) 2011-10-23T13:43:23 a hill can't change position during the game, can it? 2011-10-23T13:43:28 no 2011-10-23T13:43:39 that would be interesting though 2011-10-23T13:43:50 *** varikuzhi has quit IRC (Ping timeout: 265 seconds) 2011-10-23T13:43:54 roflmao: what's the point of A*? 2011-10-23T13:43:56 get 8 ants to surround it and carry it around 2011-10-23T13:44:08 cyphase: that would be awesome xD 2011-10-23T13:44:22 TO WAAARR!! 2011-10-23T13:44:30 *** lrh9 has joined #aichallenge 2011-10-23T13:44:46 Hello. 2011-10-23T13:45:12 hi 2011-10-23T13:45:25 or be able to spawn a queen that can spawn a new hill somewhere else 2011-10-23T13:45:26 is there a way to get the error that killed my bot? (i am using test_bot but no error is printed) 2011-10-23T13:45:32 I'm asking that because C++ starter package forgets about seen hills once it leaves sight. That's unecessary, then, since the hill will always be there. 2011-10-23T13:46:05 So I'm trying to get A* to search for the closest ant to food. How do I determine the h score if I don't actually know where the ant is going to be? Slightly confused by this >< 2011-10-23T13:50:19 bward: you have to explore until you find an ant 2011-10-23T13:50:23 tbh, i think bfs would be better 2011-10-23T13:50:51 amstan, so I can't use a heuristic function until I find an ant? 2011-10-23T13:50:59 *** levis501 has joined #aichallenge 2011-10-23T13:51:07 well, once you find an ant that isn't occupied 2011-10-23T13:51:13 bward: well, put it this way, you can't do pathfinding if you don't have a destination yet 2011-10-23T13:51:13 you're pretty much done 2011-10-23T13:51:21 bward: heuristic function of what? A* works when you have a known destination and source 2011-10-23T13:51:33 There are real problem with @tools. I try to run my C++ program and found that script sandbox.py don't sent data to my bot in line "stdin.write". This is famous problem of python version >3 (or >3.1 don't know). After calling stdin.write and stdin.flush data have not transmited to my bot. Only after calling stdin.close() i receive all data what have been sended to me. 2011-10-23T13:51:58 pedrosorio, that's how I understood it. Someone further up said I could use it to search for the nearest ant to food - I can't see how though 2011-10-23T13:52:08 In a better way that BFS would, that is. 2011-10-23T13:52:12 *** iron4o has joined #aichallenge 2011-10-23T13:52:16 python is retardedly easy to learn @_2 2011-10-23T13:52:22 bward: you can't 2011-10-23T13:52:26 @_@* I love this language 2011-10-23T13:52:27 Axoren: User error, it's not my fault. 2011-10-23T13:52:36 wut 2011-10-23T13:52:40 otzi: a hill can be destroyed while you are not "looking at it" 2011-10-23T13:52:49 pedrosorio, thanks 2011-10-23T13:53:31 once it's destroyed can't I kill the ant on top of it? 2011-10-23T13:53:44 *** dr- has joined #aichallenge 2011-10-23T13:54:01 tncardoso: you have the output from test_bot anywhere? 2011-10-23T13:54:22 otzi: what? 2011-10-23T13:54:37 ants9238483: i found some playgame options, now I am managing to get the output, thanks :) 2011-10-23T13:54:46 cool 2011-10-23T13:55:04 you destroy a hill by placing an ant on top of it, right? The ant stays there, if someone kills it, doesn't the hill reactivates? 2011-10-23T13:55:20 i think razing is permanent 2011-10-23T13:55:25 *** lllllllll has joined #aichallenge 2011-10-23T13:55:28 *** levis501 has quit IRC (Client Quit) 2011-10-23T13:55:29 otzi: have you watched any game on the website? 2011-10-23T13:55:46 no 2011-10-23T13:55:52 is someone using the clojure starter pack? 2011-10-23T13:55:55 it would be nice to have "average skill" in the latest games section 2011-10-23T13:56:09 otzi: go watch some and all your doubts will be cleared 2011-10-23T13:57:22 *** bward has quit IRC (Ping timeout: 265 seconds) 2011-10-23T13:57:29 *** ahnt has joined #aichallenge 2011-10-23T13:57:41 hi anyone here? 2011-10-23T13:58:02 Hello. 2011-10-23T13:58:29 when I upload my archive I get this message after compilation ./__MACOSX/._State.cc:1:2: error: stray '\5' in program 2011-10-23T13:58:44 which makes no sense whatsoever to me 2011-10-23T13:59:46 Me neither. 2011-10-23T14:00:27 too bad 2011-10-23T14:00:38 *** cowbandit has quit IRC (Quit: Page closed) 2011-10-23T14:00:46 *** dominater has joined #aichallenge 2011-10-23T14:01:10 hi 2011-10-23T14:02:41 *** lllllllll has quit IRC (Ping timeout: 265 seconds) 2011-10-23T14:04:10 *** Dispatch has quit IRC (Quit: Page closed) 2011-10-23T14:04:16 is someone using clojure starter package and having problem with distance? (No matching method found: abs ) 2011-10-23T14:05:25 anyone an idea what this error could mean: ./__MACOSX/._State.cc:1:2: error: stray '\5' in program 2011-10-23T14:05:43 ahnt: See my message above. 2011-10-23T14:06:16 is someone using clojure starter package and having problem with distance? (No matching method found: abs ). The contrib lib is not included in the server environment right? 2011-10-23T14:07:00 Hi bibrit - sorry the latest message I see is 13:57 2011-10-23T14:07:39 food spawns only at set coordinates, right? 2011-10-23T14:11:02 *** jesionaj has joined #aichallenge 2011-10-23T14:11:15 ahnt: I'd guess you need to delete that hidden __MACOSX folder or not add it to your zip file. 2011-10-23T14:11:49 I didn't even know I had a hidden folder in there ... I will check! 2011-10-23T14:12:03 macs love to do that 2011-10-23T14:12:05 ahnt: Some archivers add it automatically. 2011-10-23T14:12:12 aichallenge: Alexandru Stan * r5c3611f / (4 files): 2011-10-23T14:12:12 aichallenge: Merge pull request #319 from lavalamp/epsilon 2011-10-23T14:12:12 aichallenge: Updated go starter package to pay attention to hills. - http://git.io/Ewqx-A 2011-10-23T14:12:34 aichallenge: Alexandru Stan * r51a4eb7 / (2 files): 2011-10-23T14:12:34 aichallenge: Merge pull request #318 from aflag/epsilon 2011-10-23T14:12:34 aichallenge: Read player seed on setup in C++ starter package - http://git.io/nvFnpw 2011-10-23T14:12:54 ahnt: I think it's an OSX thing. That stores OSX specific meta data. 2011-10-23T14:13:12 Well it's obvious from the name it has something to do with OSX :) 2011-10-23T14:13:25 from turn#140, watch the blue ant at (8,69): http://aichallenge.org/visualizer.php?game=4165 2011-10-23T14:13:26 any idea what I can do against that? 2011-10-23T14:15:57 and I checked the fodler does not contain any hidden files 2011-10-23T14:16:08 *** deltree_ has joined #aichallenge 2011-10-23T14:16:34 I can't seem to figure out how to code a breadth first search, I keep ending up with depth first searches 2011-10-23T14:16:50 *** jesionaj has quit IRC (Quit: Leaving) 2011-10-23T14:16:54 how are you packing up your files? 2011-10-23T14:17:00 does the server allow psyco? 2011-10-23T14:17:10 ahnt: Do you not have a makefile for your project? I was under the impression that the server would use that to compile your project if available. 2011-10-23T14:17:13 *** djr_ has joined #aichallenge 2011-10-23T14:17:18 *** irchs has joined #aichallenge 2011-10-23T14:17:18 krishna: it doesn't use makefiles 2011-10-23T14:17:20 cyphase: The default start kits try to import, so I'd guess yes 2011-10-23T14:17:22 right click on the files and "compress 10 items" 2011-10-23T14:17:58 amstan: Oh. Bad assumption. :) 2011-10-23T14:18:00 hah, I think I got it 2011-10-23T14:18:07 *** bitbrit has quit IRC (Ping timeout: 260 seconds) 2011-10-23T14:18:08 krishna, that's what i thought too, but psyco doesn't work with python 2.7, which is the minimum version specified 2011-10-23T14:18:10 different type of recursion 2011-10-23T14:18:14 you might want to run zip or tar from the command line instead :/ 2011-10-23T14:18:16 I use the default cc project and include the default makefile since I only changed files and didn't add ones 2011-10-23T14:18:36 *** bitbrit has joined #aichallenge 2011-10-23T14:19:03 deltree_: use a queue 2011-10-23T14:19:17 yup, what I just figured out 2011-10-23T14:19:28 I've done them before, can't believe I didn't recognize it (it wasn't called that 2011-10-23T14:19:30 can I ask a question? 2011-10-23T14:19:41 *** Puj has joined #aichallenge 2011-10-23T14:19:47 roflmao: you've just done it 2011-10-23T14:20:25 *** ikaros_ has joined #aichallenge 2011-10-23T14:20:37 lol 2011-10-23T14:20:44 *** garlicnation has joined #aichallenge 2011-10-23T14:20:50 what are you supposed to use to prioritize squares in an a* algortihm? 2011-10-23T14:20:57 right now I'm just using the heuristic (manhattan) 2011-10-23T14:21:13 square with lowest heuristic is used 2011-10-23T14:21:19 but that leads to a lot of unnecessary squares being checked 2011-10-23T14:21:44 roflmao: i've found this page to be great for tile based a* 2011-10-23T14:21:45 http://theory.stanford.edu/~amitp/GameProgramming/Heuristics.html#breaking-ties 2011-10-23T14:22:00 *** amstan has quit IRC (Ping timeout: 240 seconds) 2011-10-23T14:22:11 *** antsyp has joined #aichallenge 2011-10-23T14:22:14 *** b4Ant has joined #aichallenge 2011-10-23T14:22:35 you are aweosme 2011-10-23T14:23:04 Does anyone have an idea why my play_game would be failing on all maps but the tutorial one? My ant works fine on the server, and on the tutorial map, but all other maps fail with no ants spawning. (I'm using all python) 2011-10-23T14:24:27 cyphase: Hah, didn't know that apparently I haven't actually been using pyco on my system all this time. 2011-10-23T14:24:28 *** lurbina has quit IRC (Read error: Connection reset by peer) 2011-10-23T14:24:40 garlicnation: can you link your game engine output? 2011-10-23T14:24:41 cyphase: Have you tried it locally? 2011-10-23T14:24:43 got the solution to my problem: was indeed some crazy thing wih zipping, the commandline zp worked 2011-10-23T14:24:47 With python 2.6 2011-10-23T14:25:24 ants9238483: http://pastebin.com/J8sHgEgW 2011-10-23T14:25:49 If someone knows how to use breakpoints (ipdb) within the sandbox (basically by replacing the sandbox by something else), please let me know 2011-10-23T14:25:50 I described the problem with tools and stdin.write here http://aichallenge.org/forums/viewtopic.php?f=25&t=1572 Does anyone can help me? 2011-10-23T14:26:12 *** ChrisH_ has joined #aichallenge 2011-10-23T14:26:18 krishna, no, i haven't tried since i remembered it wouldnt work with 2.7 2011-10-23T14:26:55 garlicnation: weird, what language are you using? 2011-10-23T14:27:37 yeah that tie breaker made a huge diff 2011-10-23T14:27:50 *** antsyp has quit IRC (Quit: Page closed) 2011-10-23T14:27:55 *** amstan has joined #aichallenge 2011-10-23T14:27:55 *** ChanServ sets mode: +o amstan 2011-10-23T14:28:24 awesome 2011-10-23T14:28:43 Belerafon_L: Shouldn't you be reading from stdin? And writing to stdout? 2011-10-23T14:28:45 ants9238483: python. I don't think the problem is in my ant code at all, since it's working on the server. I uploaded a version with my commandline string: http://pastebin.com/9jjk2n4U 2011-10-23T14:30:00 @later tell McLeopold: Here is the page source for the current rankings page when the navigation links get messed up: http://pastebin.com/tmbFpRjQ, URL was: http://aichallenge.org/rankings.php 2011-10-23T14:30:00 ChrisH_: OK 2011-10-23T14:30:56 member:ants9238483: anything other than --food none crashes the game in a different way. http://pastebin.com/LQEAbFSN 2011-10-23T14:31:27 garlicnation: you're using terminal on mac? 2011-10-23T14:31:41 *** ahnt has quit IRC (Ping timeout: 265 seconds) 2011-10-23T14:31:58 Yeah. With Zshell as my interpreter. Latest python 2.7. 2011-10-23T14:32:33 *** lrh9 has quit IRC (Quit: Leaving) 2011-10-23T14:32:38 krishna: I read from stdin on my bot, but there are no data from tools. 2011-10-23T14:32:39 Belerafon_L: Oh, I see. That's supposed to be writing to the sub process's stdin with a pipe. 2011-10-23T14:33:11 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-23T14:33:13 ants9238483: The Adobe error happens every time and doesn't have an effect on the game when it runs properly, so I don't believe that is related to the problem. 2011-10-23T14:33:49 wow my ants love commiting mass suicide 2011-10-23T14:34:03 try my play script: garlicnation: ./tools/playgame.py --fill -t 1500 --turntime=500 --loadtime=3000 --viewradius2=77 --attackradius2=5 --spawnradius2=1 --strict -l game_logs -v -m <> "$@" <> "python sample_bots/python/HunterBot.py" ...add more bots here 2011-10-23T14:34:27 Belerafon_L: Are you using an umodified starter kit bot? 2011-10-23T14:34:38 krishna: yes 2011-10-23T14:36:21 garlicnation: does it still return the same thing? 2011-10-23T14:36:23 *** pitchforkHAULR has joined #aichallenge 2011-10-23T14:39:20 *** pitchforkHAULER has quit IRC (Ping timeout: 260 seconds) 2011-10-23T14:39:21 *** pitchforkHAULR is now known as pitchforkHAULER 2011-10-23T14:40:13 ants9238483: now it doesn't even run the tutorial map. 2011-10-23T14:41:09 garlicnation: can you give me the output for that? 2011-10-23T14:41:23 basically all my ants run after the same food and end up bumping into each other and committing mass suicide 2011-10-23T14:41:53 roflmao: only assign one to each food 2011-10-23T14:41:57 yes of course 2011-10-23T14:42:12 just funny ha 2011-10-23T14:42:48 well, its usually bad to kill off all your ants just to spawn one xD 2011-10-23T14:43:09 yes xD 2011-10-23T14:43:12 member:ants9238483: ... HunterBot is crashing as far as I can tell. 2011-10-23T14:43:20 yeah hunterbot sucks ass 2011-10-23T14:43:25 *** manuranga has joined #aichallenge 2011-10-23T14:43:44 *** jhines0042 has joined #aichallenge 2011-10-23T14:43:57 afternoon 2011-10-23T14:44:06 I think I found a bug with one of the web pages. 2011-10-23T14:44:13 http://aichallenge.org/map.php?map=6&page=2 2011-10-23T14:44:16 just use the test bot then 2011-10-23T14:44:31 when viewing the map page ... you cannot seem to go to page 2 2011-10-23T14:44:36 if you're on a mac, the above script should work 2011-10-23T14:44:38 it updates the URL, but not the listings 2011-10-23T14:46:10 garlicnation: the testbot command is './tools/submission_test/TestBot.py' i think 2011-10-23T14:46:38 *** Saulzar has quit IRC (Ping timeout: 245 seconds) 2011-10-23T14:46:41 ants9238483: sorry about that. It was just a problem with HunterBot. Thanks for your help. Works great with all of my bot or if I mix TestBot in 2011-10-23T14:47:01 garlicnation: yeah, hunter does seem to fuck up constantly 2011-10-23T14:47:20 *** Macuyiko has quit IRC (Read error: Connection reset by peer) 2011-10-23T14:47:39 *** pvarga has quit IRC (Quit: pvarga) 2011-10-23T14:47:44 *** Macuyiko has joined #aichallenge 2011-10-23T14:48:25 *** pvarga has joined #aichallenge 2011-10-23T14:48:38 roflmao: any examples games for this? 2011-10-23T14:48:41 the python starter bot checks if there's less than 10ms left before ending the turn. is that really enough? 2011-10-23T14:49:05 *** dominater has quit IRC (Ping timeout: 265 seconds) 2011-10-23T14:49:08 cyphase: printing "go\n" to stdout usually doesn't take 10 ms 2011-10-23T14:49:14 actually, that's probably a lot-- yea 2011-10-23T14:50:14 *** Saulzar has joined #aichallenge 2011-10-23T14:50:56 *** Dominater has joined #aichallenge 2011-10-23T14:51:06 especially if it prints it out right there instead of exiting the function and entering another one. not sure exactly how much that takes 2011-10-23T14:51:46 Belerafon_L: Well I have no idea. I'd try different platforms and maybe python versions. 2011-10-23T14:52:06 what's the most games one person has had in 24 hours? 2011-10-23T14:52:31 I submitted my latest version 12 hours ago and i've gotten 2 games 2011-10-23T14:52:42 I have gotten 4 games in the last 24 hours 2011-10-23T14:52:53 2 with the last version 2 with the current version 2011-10-23T14:53:51 krishna: thanks. I try Python 2.7. May be 2.6? 2011-10-23T14:55:16 *** Migi32 has joined #aichallenge 2011-10-23T14:55:53 ok. I literally spent the last 20 minutes reading through the aichallenge.org site and I still don't know when the contest actually ends 2011-10-23T14:56:09 mid Decemeber 2011-10-23T14:56:17 ah ok 2011-10-23T14:56:19 at least then your final version gets used for the contest 2011-10-23T14:56:21 Migi32: see homepage 2011-10-23T14:56:42 amstan, oh... I'm stupid. 2011-10-23T14:56:44 hopefully, my bot will work now :\ 2011-10-23T14:56:55 Migi32: where were you expecting to find it? 2011-10-23T14:57:08 I checked like every other page, FAQ, specification, etc 2011-10-23T14:57:11 keeping track of states on things that might stop existing is weird 2011-10-23T14:57:28 contest rules 2011-10-23T14:57:55 if people are interested in helping to improve the faq to include more questions, please tell me 2011-10-23T14:58:03 Anyone good with python? 2011-10-23T14:58:04 I've got a syntax question 2011-10-23T14:58:08 Axoren: yep? 2011-10-23T14:58:38 I really would put it in the FAQ or the contest rules or something. For my screen's resolution, it looks like the homepage ends with the battle animation thing 2011-10-23T14:59:41 Pythoners, any idea how I could go about intercepting the KILL signal from engine.py and print a stacktrace from my bot when handled? 2011-10-23T15:00:07 SIGKILL cannot be caught. 2011-10-23T15:00:11 Mums 2011-10-23T15:00:17 I guess that's the idea.. 2011-10-23T15:00:20 amstan: You should put on the submission page that zipping or taring files on a Mac doesn't work 2011-10-23T15:00:37 Dominater: why not? it seems like it could be fixed 2011-10-23T15:00:46 SIGTERM is the polite "please clean up first" version, but SIGKILL causes the OS to immediately stop execution. 2011-10-23T15:01:19 I couln't do it last night, It kept giving me that "MyBot.*" wasn't found, tried .zip and .tar.gz 2011-10-23T15:01:30 zipped the same files on my Acer and it had no problem 2011-10-23T15:01:58 amstan: Is there some average server's speed you'll try to keep? 2011-10-23T15:02:20 UncleVasya: no, it's all based on money 2011-10-23T15:02:22 Okay, jbroman, so I can switch the signal issued to SIGTERM and then handle it in mybot? 2011-10-23T15:02:27 *** Craklyn has joined #aichallenge 2011-10-23T15:02:29 UncleVasya: but we have to concentrate more cpu power near the end 2011-10-23T15:02:32 Puj: Dunno if the engine throws SIGTERM. 2011-10-23T15:02:40 If it does, you can catch it. :P 2011-10-23T15:02:50 Nope 2011-10-23T15:02:55 But you can change it (= 2011-10-23T15:03:00 i thought this was sponsored by google? whats the point of that if you have to worry about money? 2011-10-23T15:03:02 Thanks jbroman 2011-10-23T15:03:03 :D 2011-10-23T15:03:06 *** Savaron has quit IRC (Ping timeout: 265 seconds) 2011-10-23T15:03:14 cyphase: Google does not write blank cheques. 2011-10-23T15:03:22 Well, sometimes they do :P 2011-10-23T15:03:24 Dominater: it does, you just have to zip the source file 2011-10-23T15:03:38 of course, if you have more than one file, then :P 2011-10-23T15:04:21 Uh, can anyone help me understand how to run play_one_game.cmd if I'm writing my bot in Java? 2011-10-23T15:05:22 :( 2011-10-23T15:05:30 Craklyn: compile your bot 2011-10-23T15:05:30 Craklyn: what's going wrong? 2011-10-23T15:05:44 then figure out how to execute it manually, probably something like java MyBot.class 2011-10-23T15:05:48 Well, first I want to make MyBot into a .jar file, right? 2011-10-23T15:05:56 MyBot.java -> MyBot.jar 2011-10-23T15:06:02 jbroman, one more question if you will.. Is it possible to create a stacktrace of the code currently executing immediately before the SIGTERM is handled? 2011-10-23T15:06:10 Craklyn: not really 2011-10-23T15:06:15 Craklyn: you don't necessarly need the jar, you can use the class just fine 2011-10-23T15:06:29 *** wicked has joined #aichallenge 2011-10-23T15:06:33 Craklyn: just compile it into .class, and pass 'java MyBot' to the game engine 2011-10-23T15:06:37 *** tr3v has joined #aichallenge 2011-10-23T15:06:38 *** Belerafon_L has left #aichallenge 2011-10-23T15:06:50 Puj: Yes. 2011-10-23T15:06:55 http://docs.python.org/library/signal.html 2011-10-23T15:06:55 *** wicked is now known as Guest17272 2011-10-23T15:06:58 Craklyn, You don't have to. You can use something like "java MyBot -classpath /path/to/your/classfiles" 2011-10-23T15:07:02 "The handler is called with two arguments: the signal number and the current stack frame" 2011-10-23T15:07:03 *** tr3v is now known as Guest60262 2011-10-23T15:07:12 Oooooooooooooo, thanks jbroman 2011-10-23T15:07:40 Thanks for the suggestion. I'll try this, be right back. =D =D 2011-10-23T15:07:49 nick trashant 2011-10-23T15:08:00 *** jhines0042 has quit IRC (Quit: Page closed) 2011-10-23T15:08:17 i actually couldn't get -cp to work 2011-10-23T15:08:23 @tcp 2011-10-23T15:08:24 UncleVasya: I do not know about 'tcp', but I do know about these similar topics: 'topic' 2011-10-23T15:08:32 bot crashes with no output :> 2011-10-23T15:08:41 UncleVasya: Accoun has one on the forums 2011-10-23T15:08:50 *** Guest17272 is now known as trashant 2011-10-23T15:09:00 Puj: (Better yet, since you get the full frame you can print local variables if you wish as well.) 2011-10-23T15:09:27 @topic 2011-10-23T15:09:27 ants9238483: AI Challenge (sponsored by Google): http://aichallenge.org || Channel Logs: http://contestbot.aichallenge.org || Code Repo: http://github.com/aichallenge/aichallenge || About the game (Ants): http://aichallenge.org/problem_description.php 2011-10-23T15:09:28 Thanks jbroman, hopefully that won't be necessary 2011-10-23T15:09:38 *** icefox has joined #aichallenge 2011-10-23T15:10:51 hey guys with python, i'm reading a weird thing in the tutorial: http://aichallenge.org/ants_tutorial_step_1.php 2011-10-23T15:10:57 *** Saulzar has quit IRC (Ping timeout: 258 seconds) 2011-10-23T15:11:05 amstan: I tried 2 TCP-servers yesterday, have ping to them 2500 :) 2011-10-23T15:11:06 at one point it says ants.destination, but ants is a tuple, how is this possible? 2011-10-23T15:11:12 *** Fandekasp has quit IRC (Ping timeout: 255 seconds) 2011-10-23T15:12:23 amstan: ants is the global state object thingy 2011-10-23T15:12:26 In the python code 2011-10-23T15:12:57 Okay, this is what my play_one_game file looks like: python "%~dp0playgame.py" --engine_seed 42 --player_seed 42 --end_wait=0.25 --verbose --log_dir game_logs --turns 500 --map_file "%~dp0maps\maze\maze_04p_01.map" %* "python ""%~dp0sample_bots\python\GreedyBot.py""" "python ""%~dp0sample_bots\python\LeftyBot.py""" "python ""%~dp0sample_bots\python\HunterBot.py""" "java MyBot -classpath C:\Users\Daniel\workspace\simant\bin\\" 2011-10-23T15:13:17 So basically, I only edited the very last bot to be "java MyBot -classpath c:\...." 2011-10-23T15:13:34 When I do this and run play_one_game.cmd, MyBot crashes immediately 2011-10-23T15:13:43 So I think I'm doing something wrong :( 2011-10-23T15:13:53 Craklyn: Try to build jar. 2011-10-23T15:14:14 amstan: It also holds all the utility fuctions like destination. The actual ants are in ants.ant_list 2011-10-23T15:14:26 5 minutes ago when I first asked about this, Amstan and Ants9238483 said I don't need to build the jar. 2011-10-23T15:14:33 And write in the file something like "java -jar MyBot.jar" 2011-10-23T15:14:34 krishna: can you help Axoren with it then? he has an error and i'm not exactly sure what's going on 2011-10-23T15:14:38 *** djr_ has quit IRC (Ping timeout: 252 seconds) 2011-10-23T15:15:13 Sure. 2011-10-23T15:15:57 *** sromanov has joined #aichallenge 2011-10-23T15:16:58 Craklyn, it seems like you need to use your entire package name "java -classpath C:\Users\Daniel\workspace\simant\bin\ your.package.MyBot" 2011-10-23T15:17:04 Craklyn: its all in the "setting up" step of the tutorial 2011-10-23T15:20:08 Craklyn: "play_one_game.*" only tests if the tools are working, you have to make your own "tutorial.*" to run your bot 2011-10-23T15:20:43 Okay, so I shouldn't try to do what Puj says? 2011-10-23T15:21:18 *** Dominater has quit IRC (Quit: Page closed) 2011-10-23T15:22:01 *** McLeopold has joined #aichallenge 2011-10-23T15:22:35 *** Guest60262 has quit IRC (Ping timeout: 276 seconds) 2011-10-23T15:24:33 *** CuriousSkeptic has joined #aichallenge 2011-10-23T15:26:14 Registration fails with "Could not find the email address entered. Please enter a valid email address." any site admins here that could tell me what this is about? 2011-10-23T15:27:23 if (strcmp(gethostbyname($edomain), $edomain) == 0) { $errors[] = "Could not 2011-10-23T15:27:43 CuriousSkeptic: seems like your email domain did not resolve to anything 2011-10-23T15:28:38 amstan: It's not required that the mail domain have a valid A/AAAA/CNAME record, as far as I know. 2011-10-23T15:28:53 Just common. 2011-10-23T15:28:53 *** yohann has joined #aichallenge 2011-10-23T15:28:59 hello 2011-10-23T15:29:02 jbroman: oh, you're right 2011-10-23T15:29:07 Ah, ok see the problem. milsson.nu does not exist, mx is mail.google.com 2011-10-23T15:29:17 I believe the spec is to check for an MX record, then fall back. 2011-10-23T15:29:26 i'm just starting to play with the little ants 2011-10-23T15:30:02 amstan: I can work around it if required, or will you change the test? 2011-10-23T15:30:13 CuriousSkeptic: hold on, i can revert that test 2011-10-23T15:30:40 and i don't understand why, when a ant seems to have no order, it goes up and down instead of stay quiet 2011-10-23T15:30:51 great thanx 2011-10-23T15:30:56 can anyone explain what the view radius numbers mean in ants? IE what does 55 represent? 2011-10-23T15:31:02 yohann: good luck! 2011-10-23T15:31:26 55 is a squared radius (r^2) 2011-10-23T15:31:36 peregrinator: (x1-x0)^2 + (y1-y0)^2 < viewradius2 (or is it <=?) 2011-10-23T15:31:55 You can any square is a distance-squared (as the bird flies) less than veiwradius2 2011-10-23T15:32:05 probably <= actually, yeah.. 2011-10-23T15:32:12 UncleVasya: thank you, i don't exept to win, maybe only implement a kind of djikstra to make ants go throught maze 2011-10-23T15:32:32 a1k0n: https://github.com/aichallenge/aichallenge/blob/epsilon/ants/ants.py#L1253 2011-10-23T15:32:41 Does playgame.py have a flag to forward stderr from MyBot to stderr? 2011-10-23T15:32:48 McLeopold: aha. thx. 2011-10-23T15:32:56 Puj: i dont think so 2011-10-23T15:33:02 aichallenge: Alexandru Stan * r8226c4a / website/process_registration.php : 2011-10-23T15:33:02 aichallenge: Revert "Check email for valid domain before accepting registration." 2011-10-23T15:33:02 aichallenge: This reverts commit 077050d22ecde1113d24c9d1da69966d446e3f39. 2011-10-23T15:33:02 aichallenge: Domains for emails are not required to have a valid A/AAAA/CNAME record. - http://git.io/kTQbVg 2011-10-23T15:33:03 i could also have checked my own code, since i have it open 2011-10-23T15:33:19 *** Recaiden has joined #aichallenge 2011-10-23T15:33:20 Puj: I just have a boolean in my code that i can change to either not print stderr or print it to a file 2011-10-23T15:33:30 Puj: yes, add "-e" to the command 2011-10-23T15:33:37 is there a systematic way to generate n-fold symmetric tilings in the map generator or are those hardcoded? 2011-10-23T15:33:53 Not -e? 2011-10-23T15:33:55 McLeopold 2011-10-23T15:33:57 aichallenge: Alexandru Stan * r0892257 / website/process_registration.php : removed p tags - http://git.io/QGxVFQ 2011-10-23T15:34:07 Puj: same thing 2011-10-23T15:34:14 no quotes 2011-10-23T15:34:17 CuriousSkeptic: try now 2011-10-23T15:34:26 If I under stand correct there is a rather tight timeout for a turn. Could anyone comment on the relative performance of coffee script vs, say, scala, in relation to this contest? 2011-10-23T15:34:30 amstan: will do 2011-10-23T15:34:31 heh, SymmetricMapgen.java has "gdc" instead of "gcd" 2011-10-23T15:34:39 Okay, McLeopold, but then to allow the stderr from playgame to go to terminal window, I have to disable error logging? 2011-10-23T15:35:02 no, you shouldn't have to 2011-10-23T15:35:28 it can write to the files and output to stderr/stdout all at the same time 2011-10-23T15:35:37 Hmmm, odd... All I get are my fancy 0.bot0.error and no output to terminal where I call playgame.py 2011-10-23T15:35:51 amstan: Account Activation Success! 2011-10-23T15:35:51 I think you need --verbose as well 2011-10-23T15:36:03 Omg, k tnkx mceier 2011-10-23T15:36:09 Ohops, McLeopold 2011-10-23T15:36:27 Nope, it's there McLeopold 2011-10-23T15:36:37 Yo, McLeopold! You made the Python starter, right? 2011-10-23T15:36:42 yes 2011-10-23T15:37:02 When you get a chance can I ask you some things about it? 2011-10-23T15:37:05 do you get actual output in the .error files Puj? Because that still doesn't work for me 2011-10-23T15:37:10 sure, ask now 2011-10-23T15:37:15 ants.food().remove(ants.food()[0]) 2011-10-23T15:37:19 Will that work? @_@ 2011-10-23T15:37:30 *** Ox1337 has joined #aichallenge 2011-10-23T15:37:32 *** srb has joined #aichallenge 2011-10-23T15:37:42 When I do print >> sys.stderr, "Hello" , I see it in my log files.. but not the terminal 2011-10-23T15:37:43 Axoren: it won't crash, but probably won't do what you want 2011-10-23T15:38:01 ants.food() returns a copy of the list, so you don't accidentally mess up the state 2011-10-23T15:38:05 *** Edmond has joined #aichallenge 2011-10-23T15:38:06 *** UncleVasya has quit IRC () 2011-10-23T15:38:17 McLeopold: so you can tell me why ants goes up and down when there are no issue_order() 2011-10-23T15:38:20 ? 2011-10-23T15:38:24 Axoren: do you want to remove food from futher consideration? 2011-10-23T15:38:30 *** Ox1337 has quit IRC (Client Quit) 2011-10-23T15:38:32 for the turn, yes 2011-10-23T15:38:34 *** srb has quit IRC (Client Quit) 2011-10-23T15:38:49 *** racko has quit IRC () 2011-10-23T15:38:54 yohann: you are issuing orders, check your 0.bot0.output file and check 2011-10-23T15:39:05 yohann: add -O to get that in the game logs 2011-10-23T15:39:13 McLeopold: thank you 2011-10-23T15:39:24 i try right now 2011-10-23T15:40:03 so a viewradius2 of 55 seems to suggest items within 7 tiles are viewable. But the ants seem to not actually see more than 2 tiles. What am I misunderstanding? 2011-10-23T15:40:04 Axoren: you can directly modify the ants class variables, ants.food_list. Python has no sense of private variables 2011-10-23T15:40:26 Sorry for the big paste, but McLeopold, can you see anything here which would disable MyBot stderr redirection to playgame.py stderr to terminal? 2011-10-23T15:40:27 python "%~dp0playgame.py" -So --engine_seed 42 --player_seed 42 --end_wait=0.25 --verbose --log_dir game_logs --turns 1000 --map_file "%~dp0maps\maze\maze_04p_01.map" %* "python ""%~dp0../MyBot.py""" --fill -E | java -jar visualizer.jar 2011-10-23T15:40:46 Ohhhh 2011-10-23T15:40:49 lol 2011-10-23T15:40:58 I guess the visualizer has to redirect the stderr as well 2011-10-23T15:41:04 :) 2011-10-23T15:41:06 Okay, I was advised to work through the tutorial (http://aichallenge.org/ants_tutorial.php) to understand how to run a java bot locally using play_one_game.cmd... When I run the step where I run "make", I get the message "javac: Command not found". I don't see any package called javac to install for cygwin. Any suggestions? 2011-10-23T15:41:08 Winning. 2011-10-23T15:41:31 Craklyn: windows or linux? 2011-10-23T15:41:37 cygwin, so linux on windows :P 2011-10-23T15:42:01 check your path, you should be able to run javac by itself and get help 2011-10-23T15:42:02 Craklyn: if you're running windows, you need to install jdk 2011-10-23T15:42:24 so it would be ants.food_list.remove(ants.food()[0])? 2011-10-23T15:42:25 Ah, so if I install jdk, it'll be usable by cygwin? 2011-10-23T15:42:35 Right, McLeopold? 2011-10-23T15:42:36 No worries if you're not sure :P 2011-10-23T15:42:43 Craklyn: type "which javac" to see if it is installed and where 2011-10-23T15:43:12 It's not installed yet 2011-10-23T15:43:27 Axoren: yes, but I wouldn't do that. I think it is better to maintain a list of food locations you have targetted, then check that list, rather than remove stuff from the state 2011-10-23T15:43:48 but that only affects the current turn, correct? 2011-10-23T15:44:08 Axoren: also, you are removing the first item, so ants.food_list.pop() would look nicer 2011-10-23T15:44:14 alright 2011-10-23T15:44:23 Axoren: yes, it will heal itself next turn 2011-10-23T15:44:39 alright 2011-10-23T15:44:40 Axoren: you just open yourself up for bugs if you forget it is happening 2011-10-23T15:44:52 so for the moment, it'll suffice 2011-10-23T15:45:05 Ducktape all the things! 2011-10-23T15:45:15 Axoren: actually, what is better is to get a list calling ants.food(), then remove from that temporary list 2011-10-23T15:45:18 Craklyn: jdk should work with cygwin 2011-10-23T15:45:27 Craklyn: but not 100% positive 2011-10-23T15:45:29 food = ants.food() food.pop() 2011-10-23T15:45:35 *** cirno_the_greate has joined #aichallenge 2011-10-23T15:45:46 Okay, thanks :) 2011-10-23T15:45:56 Puj, with the code you pasted, do you see the "print >> sys.stderr, "Hello"" stuff in your logfiles? I can't get that to work for the life of jesus 2011-10-23T15:46:02 Actually did that as you typed it, McLeo 2011-10-23T15:46:05 Craklyn: http://cs.nyu.edu/~yap/prog/cygwin/FAQs.html 2011-10-23T15:46:23 *** jezuz` has joined #aichallenge 2011-10-23T15:46:26 My code is lookling less like a jamaican bobsled team in Cancun 2011-10-23T15:46:50 Axoren: just pop it in the freezer 2011-10-23T15:47:05 One more question, more about python than about ants 2011-10-23T15:47:31 I'm still a little confused about viewradius and ant vision: a viewradius2 of 55 seems to suggest items within 7 tiles are viewable. But the ants seem to not actually see more than 2 tiles. What am I misunderstanding? 2011-10-23T15:47:41 If I have a list of tuples, can I modify the contents of the tuples without removing them from the list and placing new tuples in the list? 2011-10-23T15:47:49 or do I have to remove then replace? 2011-10-23T15:48:21 Any administrators around that can help me out with a code uploading problem? My bot runs fine locally with the tools test and game playing, but fails when the server runs its test, midway through. The error output is half a useful traceback at the bottom, and valid output at the top, truncated with no way to view the whole log 2011-10-23T15:48:35 Most of the starter packages mention "hills not implemented yet". What does it mean to implement hills? 2011-10-23T15:48:44 Axoren: tuples are immutable 2011-10-23T15:48:52 ugh 2011-10-23T15:48:56 tuples are not meant to be changed, use a list instead 2011-10-23T15:49:03 *** pitchforkHAULR has joined #aichallenge 2011-10-23T15:49:18 [] then? 2011-10-23T15:49:24 yeah 2011-10-23T15:49:30 or, remove the tuple and add a new one to the list 2011-10-23T15:49:49 now, lets say I had x = [a, [b, c]] and I wanted to change b to new b 2011-10-23T15:49:57 x[1][0] = new b 2011-10-23T15:49:57 CuriousSkeptic: in the ruby started package it meant it didn't parse anything when the game sent it a location that had a hill (h 10, 10 for instance) 2011-10-23T15:49:58 right? 2011-10-23T15:50:09 Axoren: Yes. 2011-10-23T15:50:16 *** jezuz has quit IRC (Ping timeout: 260 seconds) 2011-10-23T15:50:20 SWEET JESUS! I'M LEARNING! 2011-10-23T15:50:36 Ah, ok, so it's just a matter of some tweaking to get it working then 2011-10-23T15:50:51 *** g0llum has quit IRC (Read error: Connection reset by peer) 2011-10-23T15:50:54 *** wifiget has joined #aichallenge 2011-10-23T15:51:01 *** lurbina has joined #aichallenge 2011-10-23T15:51:20 CuriousSkeptic: yeah you'll eventually want some code to send ants to 'raze' those hills 2011-10-23T15:52:01 *** pitchforkHAULER has quit IRC (Ping timeout: 260 seconds) 2011-10-23T15:52:01 *** pitchforkHAULR is now known as pitchforkHAULER 2011-10-23T15:52:06 {} denotes a dictionary/map, right? 2011-10-23T15:52:31 *** Accoun has quit IRC () 2011-10-23T15:52:37 Axoren: Yes. 2011-10-23T15:52:44 { key: value, key2: value2 } 2011-10-23T15:52:52 boegel: in direction p1 is destination? 2011-10-23T15:54:01 *** McLeopold has quit IRC (Ping timeout: 244 seconds) 2011-10-23T15:54:28 boegel: navigation on torus works 2011-10-23T15:54:45 I looked the log, so if i understand well, my bot issue_order to go in the water, and the ant go opposite side 2011-10-23T15:54:59 is that a normal behavior? 2011-10-23T15:55:15 *** McLeopold has joined #aichallenge 2011-10-23T15:55:36 cirno_the_greate: no, p2 is 2011-10-23T15:55:43 so can anyone explain how far ants can 'see' 2011-10-23T15:55:52 cirno_the_greate: you look into which direction you would go from p1 to p2 2011-10-23T15:56:00 hmm 2011-10-23T15:56:00 *** manuranga has quit IRC (Quit: Page closed) 2011-10-23T15:56:03 I looked the log, so if i understand well, my bot issue_order to go in the water, and the ant go opposite side, is it normal ? 2011-10-23T15:56:03 cirno_the_greate: you tested it with your bot? 2011-10-23T15:56:10 *** tr3v has joined #aichallenge 2011-10-23T15:56:11 boegel: i'm testing it in repl 2011-10-23T15:56:14 peregrinator: you get a parameter "viewradius2" at the start, that is the view distance squared 2011-10-23T15:56:20 *** lericson has joined #aichallenge 2011-10-23T15:56:24 and checking coordinates on visualisator 2011-10-23T15:56:29 Hmm, my account lericson did not receive an activation e-mail. 2011-10-23T15:56:29 cirno_the_greate: ok 2011-10-23T15:56:35 *** tr3v is now known as Guest69304 2011-10-23T15:56:41 And yes I did check my spam filter. 2011-10-23T15:56:43 cirno_the_greate: I haven't checked my implementation, but I think it's right 2011-10-23T15:56:52 cirno_the_greate: if I got the x,y coords right, that is 2011-10-23T15:56:57 peregrinator: so if viewradius2 = 77 (as it is on the server), your ants can see ~8,78 squares around them 2011-10-23T15:57:17 boegel: after checking _my_ implementation i reversed all signs ;) 2011-10-23T15:57:25 *** deltree_ has quit IRC (Quit: If you don't like it, don't suck at it) 2011-10-23T15:57:30 this is a little bit tricky stuff 2011-10-23T15:57:54 if you swap p1 and p2 - everything works as intended 2011-10-23T15:57:56 olexs1 - ok - so when I run the tutorial map it's set to 55 2011-10-23T15:57:56 Who do I talk to if my account is unusable? 2011-10-23T15:57:59 yohann: I think your bot is confused, or you are, that is not normal 2011-10-23T15:58:00 check it please :) 2011-10-23T15:58:03 cirno_the_greate: I think it should be right, but haven't checked 2011-10-23T15:58:09 yep, i know that 2011-10-23T15:58:09 *** yohann has quit IRC (Quit: Lost terminal) 2011-10-23T15:58:17 cirno_the_greate: I don't have a bot to check it with :P 2011-10-23T15:58:25 olexs1: but the ants only seem to see 2 squares rather than ~7.4 2011-10-23T15:58:37 hm 2011-10-23T15:58:46 okay, i'll check it with bot 2011-10-23T15:58:50 peregrinator: that can't be right then 2011-10-23T15:59:31 peregrinator: 2 squares sounds more like the attack radius, which is just under 2 (sqrt(5) iirc) 2011-10-23T16:00:03 peregrinator: until sometime yesterday the tools did have the wrong default viewradius2 (55 instead of 77) 2011-10-23T16:00:08 *** Craklyn has quit IRC (Ping timeout: 265 seconds) 2011-10-23T16:00:17 olexs1 - the first turn it seems to be 7 squares (I get 4 food reported) but the second turn I only get 3 reported (a 2 square radius) 2011-10-23T16:00:23 cirno_the_greate: hmm, seems like I have things reversed, you're right 2011-10-23T16:00:32 you can look in playgame.py to check what your's is, or just download the tools again 2011-10-23T16:00:42 *** Craklyn has joined #aichallenge 2011-10-23T16:00:45 janzert olexs1: I'll change the default radius 2011-10-23T16:00:55 peregrinator: I'm not sure I am 100% correct, but I think you only get food reported once, when it first comes into view 2011-10-23T16:01:09 food is reported every turn it is visible 2011-10-23T16:01:12 as are hills 2011-10-23T16:01:20 okay, then I remembered wrong 2011-10-23T16:01:22 and ants :) 2011-10-23T16:01:25 just not water 2011-10-23T16:01:38 cirno_the_greate: it is correct in the horizontal direction though? 2011-10-23T16:01:45 cirno_the_greate: I just swapped North/South 2011-10-23T16:02:12 janzert: out of curiosity, did you get pinged when i mentioned you in that ada pull request? 2011-10-23T16:02:28 boegel: no, swap for both directions 2011-10-23T16:02:46 I have this guy from this channel 2011-10-23T16:02:51 asking me if I'm on a dating site o.o 2011-10-23T16:02:55 amstan: looks like I did 2011-10-23T16:03:01 cool 2011-10-23T16:03:02 just haven't looked at my email since then 2011-10-23T16:03:10 janzert: also, i have no idea why he closed it 2011-10-23T16:03:15 cirno_the_greate: hmm 2011-10-23T16:03:47 janzert: odd because the first turn I get f 7 19, f 14 17, f 14 21, f 16 19 reported 2011-10-23T16:03:49 *** irchs has quit IRC (Quit: irchs) 2011-10-23T16:04:07 janzert: but the second turn the f 7 19 is no longer reported 2011-10-23T16:04:16 janzert: as if I can't see it 2011-10-23T16:04:19 cirno_the_greate: I don't see how I got that wrong... 2011-10-23T16:05:03 *** Recaiden has quit IRC (Quit: Page closed) 2011-10-23T16:05:03 janzert: for reference the ant is at 14, 19 - is this some kind of bug? 2011-10-23T16:05:06 boegel: 1 moment, may be i was wrong; attaching it to my bot atm 2011-10-23T16:05:11 cirno_the_greate: here's what I have now, this is correct imho http://hpaste.org/53032 2011-10-23T16:05:36 peregrinator: are you sure you don't step out of view? can you paste the replay at paste.aichallenge.org? 2011-10-23T16:05:49 *** wifiget is now known as csyncope 2011-10-23T16:06:26 janzert: amstan: is there a way to check if our emails get bounced or rejected? 2011-10-23T16:06:45 McLeopold: check postfix logs? 2011-10-23T16:06:50 my password on fluxid is really ridiculous to remember... I didn't think I'd have to remember it 2011-10-23T16:06:54 McLeopold: not sure what you can grep for though 2011-10-23T16:07:01 where is it? 2011-10-23T16:07:03 it's like 25 characters of my just hitting my keyboard xD 2011-10-23T16:07:03 Bounces should be getting sent to staff@aichallenge.org 2011-10-23T16:07:07 postfix log won't have it 2011-10-23T16:07:20 but I think the google group is eating them :( 2011-10-23T16:07:22 *** icefox has quit IRC (Quit: icefox) 2011-10-23T16:07:30 janzert: http://paste.aichallenge.org/oBCIM/ 2011-10-23T16:07:57 I was trying to figure out the other day if there was some way in the google group to find them 2011-10-23T16:08:23 viewradius is 55 in that replay 2011-10-23T16:08:41 janzert: nm, I am dumb - somehow I misread my visualizer and didn't see the ant move down one square 2011-10-23T16:09:02 oh man! normalizer for bayes rule! so easy! 2011-10-23T16:09:05 peregrinator: glad it's figured out 2011-10-23T16:09:20 amstan: can you get McLeopold added to the staff@aichallenge.org? 2011-10-23T16:09:24 janzert: me too - couldn't figure out why my math made no sense! 2011-10-23T16:09:28 janzert: sure 2011-10-23T16:09:32 amstan: do you mean the "eta" trick from aiclass.com? 2011-10-23T16:09:39 boegel: yes 2011-10-23T16:09:49 amstan: yeah, it rules :) 2011-10-23T16:10:04 when you call ants.getFoodTiles is it only the food that the ant can see or all food tiles on the board 2011-10-23T16:10:24 *** pitchforkHAULER has quit IRC (Quit: - nbs-irc 2.39 - www.nbs-irc.net -) 2011-10-23T16:10:41 janzert: staff is set to All users within aichallenge.org 2011-10-23T16:10:42 Can someone tell me if I'm just being python-tarded or if I've got another minor coding error? 2011-10-23T16:10:45 amstan: any idea if there is a place we can see what bounce messages are ending up at at that address too? 2011-10-23T16:10:48 My bot 2011-10-23T16:10:56 does McLeopold have an account? 2011-10-23T16:11:09 janzert: nope 2011-10-23T16:11:11 s timing out at turn 1 2011-10-23T16:11:27 *** Accoun has joined #aichallenge 2011-10-23T16:11:36 I think I can get the bounces sent to me personally, if we can't get the list to handle them 2011-10-23T16:11:45 cirno_the_greate: so, does it work as expected? 2011-10-23T16:11:57 *** lurker_for_life has joined #aichallenge 2011-10-23T16:12:22 *** CuriousSkeptic has quit IRC (Quit: Page closed) 2011-10-23T16:12:46 Anyone care to look over my code and help me find out what's wrong? 2011-10-23T16:12:55 *** vik has joined #aichallenge 2011-10-23T16:13:05 *** djr_ has joined #aichallenge 2011-10-23T16:14:07 *** vik has quit IRC (Client Quit) 2011-10-23T16:14:21 looks like I can set the ReturnPath header on the outgoing emails so the bounces will go directly to me 2011-10-23T16:14:37 but people replying will still get staff@aichallenge 2011-10-23T16:14:40 boegel: still working with code 2011-10-23T16:14:48 i need to refactor everything from the ground 2011-10-23T16:15:00 ._. 2011-10-23T16:15:05 cirno_the_greate: ok 2011-10-23T16:15:21 cirno_the_greate: I've pushed the code to my github repo, I think it's fine now 2011-10-23T16:15:30 boegel: yep 2011-10-23T16:15:37 cirno_the_greate: pull it if you think it's ok, or letme know what's wrong 2011-10-23T16:15:43 cirno_the_greate: I have to run now, ttyl 2011-10-23T16:15:45 sure 2011-10-23T16:15:49 boegel: okay 2011-10-23T16:15:53 i'll just pull then 2011-10-23T16:15:59 and apply fixes if needed 2011-10-23T16:16:11 Hi again- I'm trying to work through the tutorial (http://ai-contest.com/ants_tutorial.php) for java using cygwin. When I run ./tutorial.sh, I get a crash: "Could not find the main class: MyBot. Program will exit." 2011-10-23T16:16:44 Craklyn: did you run make? 2011-10-23T16:16:47 yes 2011-10-23T16:16:53 do you have a jar file? 2011-10-23T16:16:56 yep 2011-10-23T16:17:20 can you run "java -jar MyBot.jar" by itself, enter "ready" and get back a "go"? 2011-10-23T16:17:33 When I run that, I get the exact same error 2011-10-23T16:20:31 I tried making "tutorial.cmd" and running it on windows. This actually did work... 2011-10-23T16:21:01 Craklyn, how did you make the jar file? 2011-10-23T16:21:17 Yes, I have MyBot.jar in the directory after I typed "make" 2011-10-23T16:21:43 *** stano has joined #aichallenge 2011-10-23T16:21:44 What command did you use to build your jar file... 2011-10-23T16:21:50 "make" 2011-10-23T16:22:09 Craklyn, You have a file named MyBot.java and the class in that file is named MyBot, right? 2011-10-23T16:22:09 There's a file 'Makefile' included in the download package 2011-10-23T16:22:41 Yes, it's the unmodified MyBot.java code, which contains the class MyBot 2011-10-23T16:22:51 *** AlvinMinring_ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T16:23:19 A* is too slow apparently 2011-10-23T16:23:32 which is bull 2011-10-23T16:23:36 does A* work well with boundaries? 2011-10-23T16:23:37 (And inside MyBot.java there's a "public static void main" function...) 2011-10-23T16:23:53 Anyone have any tips for localizing timeout problems in python? 2011-10-23T16:23:58 *** xathis_ has joined #aichallenge 2011-10-23T16:24:15 aichallenge: janzert * r226e8f1 / website/email.php : Add ReturnPath to registration emails - http://git.io/mQQVJw 2011-10-23T16:24:19 Puj: run with python -m cProfile MyBot.py < 0.bot0.input 2011-10-23T16:24:48 *** xathis has quit IRC (Ping timeout: 240 seconds) 2011-10-23T16:25:22 Craklyn: try this command: javac MyBot.java 2011-10-23T16:25:41 When I enter that, I get no text output, but it runs fine 2011-10-23T16:25:48 then do python tools/playgame.py "java MyBot" "python tools/sample_bots/python/HunterBot.py" --map_file tools/maps/example/tutorial1.map --log_dir game_logs --turns 60 --scenario --food none --player_seed 7 --verbose -e 2011-10-23T16:25:52 McLeopold, thanks m8 2011-10-23T16:26:18 cProfile is good stuff 2011-10-23T16:26:47 aichallenge: janzert * ra2c7e41 / website/email.php : Add missing ) - http://git.io/6E9QFw 2011-10-23T16:26:47 Craklyn: I get the same issue with cygwin 2011-10-23T16:26:52 Craklyn: still have the no class error? 2011-10-23T16:27:34 oh, Axoren, hi 2011-10-23T16:27:49 yo 2011-10-23T16:27:49 I'm at work and I'm pissed, hold on 2011-10-23T16:28:16 Errmm, now I have a messier error: "0 [main] python 3412 C:\cygwin\bin\python.exe: ** fatal error - unable to remap \\?\C:\cygwin\lib\python2.6\lib-dynload\select.dll to same address as parent: 0x360000 != 0x3F0000 (Then it gives a stack trade).. 2011-10-23T16:28:32 Exit cygwin processes 2011-10-23T16:28:38 Start cmd 2011-10-23T16:28:50 Go to /bin 2011-10-23T16:28:53 Type 'ash' 2011-10-23T16:29:02 Axoren, I have a little something for ya to lighten your day >:P 2011-10-23T16:29:04 < Axoren> draw dicks with my ants when food and enemies aren't present 2011-10-23T16:29:10 Axoren, http://ash.webfactional.com/replay.610 2011-10-23T16:29:18 did you really? 2011-10-23T16:29:29 *whistles innocently* 2011-10-23T16:29:32 Type '/bin/rebaseall' 2011-10-23T16:30:02 Then close the cmd prompt and try again 2011-10-23T16:30:24 Axoren, you can't blame me, it was partially your idea! :D 2011-10-23T16:30:34 I love you 2011-10-23T16:30:36 So much 2011-10-23T16:30:41 *** mpan has joined #aichallenge 2011-10-23T16:30:45 you should make more than one 2011-10-23T16:30:51 and have that sword fight the other guy was talking about 2011-10-23T16:31:07 I got this in response: /usr/bin/tclpip84.dll: skipped because not rebaseable. rebase: failed to create temporary rebase database: No such file or directory 2011-10-23T16:31:18 I'm still having issues with my pathfinding algorithm taking too long to be of any use 2011-10-23T16:31:24 *** djs has quit IRC (Quit: Page closed) 2011-10-23T16:31:28 aichallenge: McLeopold * rd71f3d5 / website/starter_packages.php : add tools link to starter packages page - http://git.io/keHoww 2011-10-23T16:31:32 Axoren, hehe :P 2011-10-23T16:31:40 I need to find a way to partition goals into mini goals 2011-10-23T16:31:59 because finding THE BEST is just leaving me in the dust 2011-10-23T16:32:09 *** thestinger has left #aichallenge ("WeeChat 0.3.6") 2011-10-23T16:33:26 Craklyn: TBH i would just stick with the native windows environment if it's not too much trouble 2011-10-23T16:33:48 *** Palmik has quit IRC (Remote host closed the connection) 2011-10-23T16:33:59 I wanted to avoid it because I'm much worse at using \ than /. But I think you might be right :) 2011-10-23T16:34:36 aichallenge: janzert * r1f718b4 / website/starter_packages.php : Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://git.io/pG6G7Q 2011-10-23T16:34:38 aichallenge: janzert * rc36be5c / website/lookup.php : Correct variable name - http://git.io/4Cgcag 2011-10-23T16:34:43 *** irchs has joined #aichallenge 2011-10-23T16:35:13 *** mpan has quit IRC (Quit: Page closed) 2011-10-23T16:35:44 amstan: the java makefile doesn't add all the classes, just the mybot.class? 2011-10-23T16:35:52 *** delt0r_ has quit IRC (Ping timeout: 244 seconds) 2011-10-23T16:35:54 McLeopold: idk 2011-10-23T16:36:10 didn't you write it 2011-10-23T16:36:27 CLASSES needs to be all class files, it looks like it is just MyBot.class 2011-10-23T16:37:11 Puj / ants9238483 - Thanks a lot for your help :) 2011-10-23T16:37:22 McLeopold: java doesn't pull them in automatically? 2011-10-23T16:37:28 no prob 2011-10-23T16:37:31 I don't really know 2011-10-23T16:38:01 hmmm 2011-10-23T16:38:11 does php's json_decode truncate long output? 2011-10-23T16:38:20 aichallenge: McLeopold * r77cdcc2 / ants/dist/starter_bots/java/Makefile : fix java bot make file issue - http://git.io/eWf_Fw 2011-10-23T16:38:21 aichallenge: McLeopold * r5941333 / website/lookup.php : Merge branch 'epsilon' of https://github.com/aichallenge/aichallenge into epsilon - http://git.io/IGfwFA 2011-10-23T16:38:25 slow 2011-10-23T16:38:33 *** daniloresende has joined #aichallenge 2011-10-23T16:38:35 I'm not seeing any truncation happening in the worker's testing of new compiles 2011-10-23T16:38:39 in this game ant cant stay on plase? 2011-10-23T16:38:55 *** bmh has joined #aichallenge 2011-10-23T16:39:18 anyone else here using haskell? 2011-10-23T16:39:41 Craklyn: it was a Makefile issue, you can pull the new starter pack 2011-10-23T16:40:43 Aha, thanks very much McLeopold :) 2011-10-23T16:42:14 amstan: okay, I think my Makefile fix is still broken 2011-10-23T16:42:29 ergh, amstan don't go reverting stuff just cause someone in the issue tracker wants it done :P 2011-10-23T16:42:45 janzert: the mail thing? 2011-10-23T16:43:05 yeah, it was just using the wrong function is all 2011-10-23T16:43:14 *** boegel has quit IRC (Quit: Leaving) 2011-10-23T16:43:16 should be getmxrr instead of gethostbyname 2011-10-23T16:43:17 janzert: it wasn't based on issue tracker, it was based on a user waiting to register with a valid email address 2011-10-23T16:43:36 and that check is catching a lot of bad email adresses 2011-10-23T16:44:01 ah 2011-10-23T16:44:08 McLeopold: revert that commit, add your other files to the SOURCES= list 2011-10-23T16:44:11 McLeopold: space separated 2011-10-23T16:45:40 *** Edmond has quit IRC (Quit: Page closed) 2011-10-23T16:46:02 aichallenge: McLeopold * rb922a4d / ants/dist/starter_bots/java/Makefile : fix java bot make file issue - http://git.io/dwIEVw 2011-10-23T16:46:44 *** djr_ has quit IRC (Ping timeout: 260 seconds) 2011-10-23T16:47:12 You can order 2 ants to the same space. If you do this, both ants will die. You can be on the same space as an enemy. This can happen if an ant spawns next to an enemy. (You might also be close to losing your hill!) 2011-10-23T16:47:23 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T16:47:24 I'm confused by the second case 2011-10-23T16:48:36 *** delt0r_ has joined #aichallenge 2011-10-23T16:48:42 *** Macuyiko has quit IRC (Ping timeout: 252 seconds) 2011-10-23T16:50:12 aichallenge: McLeopold * rc7476b1 / ants/dist/starter_bots/java/Makefile : fix java bot make file issue - http://git.io/gpx8fw 2011-10-23T16:50:13 *** cirno_the_greate has quit IRC (Remote host closed the connection) 2011-10-23T16:50:42 *** lurbina has quit IRC (Quit: Leaving) 2011-10-23T16:51:09 aichallenge: janzert * r1e07e03 / website/process_registration.php : Add corrected check for email domain, Fix #322 - http://git.io/uD8GLQ 2011-10-23T16:51:47 rmmh: ants fight by colliding, 'eh? 2011-10-23T16:51:57 *** ltriant has joined #aichallenge 2011-10-23T16:52:05 no, how you could have an ant on the same square as an enemy 2011-10-23T16:52:13 rmmh: I'll rewrite it 2011-10-23T16:52:23 oh wait 2011-10-23T16:52:30 yeah I don't see it 2011-10-23T16:52:34 *** djr_ has joined #aichallenge 2011-10-23T16:52:41 looks like a residual from when ants spawned directly from food 2011-10-23T16:52:43 by having an ant spawn on a hill as an enemy also steps on it 2011-10-23T16:52:55 janzert: phases are move -> attack -> raze -> gather -> spawn 2011-10-23T16:52:59 uh. Why is Zannick banned from this channel? 2011-10-23T16:53:08 wtf? 2011-10-23T16:53:13 shouldn't that mean that raze should happen before spawn? 2011-10-23T16:53:13 that's not me banned 2011-10-23T16:53:18 Zannick: oh :) 2011-10-23T16:53:19 that's a ban i set on someone else 2011-10-23T16:53:27 Zamick? 2011-10-23T16:53:28 rmmh, turns are computed in 5 phases: (in order) move, attack, raze, gather, spawn. normally you and an enemy cant be on the same space because someone would have been killed before you got that close. however, if in one turn an enemy moves to a space next to your hill and your hill spawns an ant, then on the next turn your ant moves onto that enemy's space, you'll be on the same space until the attack portion of the turn is completed 2011-10-23T16:53:42 * bmh fails reading comprehension 2011-10-23T16:53:48 ahhh 2011-10-23T16:53:51 *** wnight has joined #aichallenge 2011-10-23T16:53:52 i know you knew some of that, but i had already typed it :) 2011-10-23T16:53:53 Zannick: what did he do? 2011-10-23T16:53:58 Zannick: my bot will kick your bot's ass :) 2011-10-23T16:54:05 bmh: i'm an admin 2011-10-23T16:54:05 * bmh hasn't written a bot yet... 2011-10-23T16:54:11 *** avers_ has joined #aichallenge 2011-10-23T16:54:27 Zannick: thanks. I owe you a (free) beer. 2011-10-23T16:54:28 cyphase: but it's impossible to have two ants on the same square after all the phases right 2011-10-23T16:54:29 The Ruby starter kit I downloaded doesn't recognize 'h x y z' commands. Looking at the java starter kit that appears to be Anthills. 2011-10-23T16:54:42 rmmh: read it again, I changed the wording 2011-10-23T16:54:51 *** Areks has quit IRC (Ping timeout: 252 seconds) 2011-10-23T16:54:57 rmmh, yes, i believe so 2011-10-23T16:55:05 Zannick: oh wow, he's a special type of person 2011-10-23T16:55:10 amstan: yep. 2011-10-23T16:55:14 McLeopold: okay 2011-10-23T16:56:01 Alexer! 2011-10-23T16:56:05 Draw more penises! 2011-10-23T16:56:34 definitely not someone who should be here 2011-10-23T16:56:35 Axoren :D 2011-10-23T16:56:37 rmmh, which section was that in? 2011-10-23T16:56:49 cyphase: "Collisions" 2011-10-23T16:56:58 Is the ruby starter kit out of date? It doesn't recognize h or player_seed ("unexpected: h 1 18 0", etc) 2011-10-23T16:57:31 Zannick: I'm guessing you know a1k0n? (I haven't met him) 2011-10-23T16:57:35 ah, there it is. thanks 2011-10-23T16:58:02 *** chezz has joined #aichallenge 2011-10-23T16:58:02 bmh: i haven't met him either 2011-10-23T16:58:18 McLeopold: collisions happen during the move stage, right? 2011-10-23T16:58:42 McLeopold, so the engine checks enemy collisions before checking attack radius? 2011-10-23T16:59:25 McLeopold: any idea how user 2839 ended up with a country_id of 999? 2011-10-23T17:00:50 hmm, actually the registration just checks that country code is not empty 2011-10-23T17:00:55 not that it is valid 2011-10-23T17:01:09 shucks, score isn't available during the game, 'eh? 2011-10-23T17:01:29 Nope 2011-10-23T17:01:45 the link on 2839's profile page doesn't have an id in the country link 2011-10-23T17:03:55 *** McLeopold has left #aichallenge 2011-10-23T17:05:53 aichallenge: janzert * rd2d2109 / website/process_registration.php : Add check for valid country to process_registration - http://git.io/QhbdxA 2011-10-23T17:07:08 amstan: there must be a link to aichallenge.org/forum somewhere, lot of people hitting it anyway 2011-10-23T17:07:27 :q 2011-10-23T17:07:46 *** pvarga_ has joined #aichallenge 2011-10-23T17:07:48 Zannick: to drag crawl into this -- the game should use squarelos ;) 2011-10-23T17:08:15 janzert: not on our main site 2011-10-23T17:08:20 janzert: we should get a permanent redirect 2011-10-23T17:08:32 bmh: square line of sight? 2011-10-23T17:08:39 posts from the forums might have it 2011-10-23T17:08:52 rmmh: yeah, like the king in chess 2011-10-23T17:09:18 circular LOS is strange when you only have four directional movement 2011-10-23T17:09:32 bmh: ಠ_ಠ 2011-10-23T17:09:33 You mean it's awesome? 2011-10-23T17:09:38 I like circular LOS :) 2011-10-23T17:09:40 Zannick: <3 2011-10-23T17:09:45 Dem eyes 2011-10-23T17:10:24 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-23T17:10:24 *** pvarga_ is now known as pvarga 2011-10-23T17:10:51 bmh: we shouldn't have sight at all, we should receive scent inputs from each ant 2011-10-23T17:10:57 *** chezz has quit IRC (Quit: Page closed) 2011-10-23T17:11:01 rmmh: for serious 2011-10-23T17:11:08 Okay, so when I try to run make in the windows commandline, I get the message "Could not find C:\Users\Daniel\workspace\aichallenge\*.class 2011-10-23T17:11:33 The .jar file is created, but since it has that problem, it halts any script which runs make along the way 2011-10-23T17:11:37 *** tncardoso has quit IRC (Quit: bye) 2011-10-23T17:11:42 where strenght of each scent is sum of 1/d^2 (manhattan, of course) 2011-10-23T17:11:50 cookies need love like people do 2011-10-23T17:11:52 I don't know much about .cmd scripts/windows command line, so I'm not sure what to do :( 2011-10-23T17:11:55 Craklyn, are your class files there or in like ./bin? 2011-10-23T17:12:00 om nom nom that's why I'm a cannibal 2011-10-23T17:12:06 The class files are in /bin 2011-10-23T17:12:11 ../bin rather 2011-10-23T17:12:34 Okay, well, first try copying your class files there, then make the jar 2011-10-23T17:13:07 *** Mixailo has quit IRC (Ping timeout: 265 seconds) 2011-10-23T17:16:28 *** Novarg has joined #aichallenge 2011-10-23T17:17:15 I feel like clever symmetry detection could be a big advantage 2011-10-23T17:17:16 *** cybsy has quit IRC (Ping timeout: 256 seconds) 2011-10-23T17:17:43 *** EspenG has quit IRC (Quit: Leaving) 2011-10-23T17:18:00 *** pvarga has quit IRC (Ping timeout: 240 seconds) 2011-10-23T17:20:06 Okay, I did that. When I run "make" inside my tutorial.cmd, the tutorial script still halts after "make" for some reason :( 2011-10-23T17:20:35 *** EspenG has joined #aichallenge 2011-10-23T17:21:25 *** cybsy has joined #aichallenge 2011-10-23T17:21:58 and you tried doing the commands in the .cmd manually via the command line, Craklyn? 2011-10-23T17:23:02 Anyone know what would be causing "python -m cProfile ../MyBot.py < game_logs/0.stream" to become "python -m cProfile ../MyBot.py 0 Yeah, each step appears to work. The last steps are "Make / cd ../../aichallenge / python tools/playgame.py ...." 2011-10-23T17:23:30 The script halts after "Make" 2011-10-23T17:23:52 And when I run Make.cmd by itself without turning off the Echo, I see no errors or other reasons it should stop 2011-10-23T17:25:19 I dunno Craklyn... Sorry man.. 2011-10-23T17:25:27 *** pvarga has joined #aichallenge 2011-10-23T17:26:05 No worries :P Maybe I should work on this on my macbook instead, since I don't have any idea how to run scripts in windows :P 2011-10-23T17:27:01 By the way, any word if the servers are going to be more responsive in the future? I was just testing my program there, until my programs stopped executing entirely today 2011-10-23T17:27:23 *** azEf has quit IRC () 2011-10-23T17:27:25 Have you tried the tcp server or is that what you mean? 2011-10-23T17:27:52 *** EspenG has quit IRC (Quit: Leaving) 2011-10-23T17:28:07 I meant the ai-challenge.com page. I did try the tcp server, and got some error. So I was trying to understand how to run the game locally so I could maybe understand how to use the tcp server....haha 2011-10-23T17:28:15 I guess I've come full circle today... 2011-10-23T17:28:45 lol, A whole lot of learning it seems (= 2011-10-23T17:28:57 *** Guest69304 has quit IRC (Ping timeout: 252 seconds) 2011-10-23T17:31:31 *** dinosoep has joined #aichallenge 2011-10-23T17:31:34 For the tcp servers, the script uses a username and password. What username and password do they want? 2011-10-23T17:31:37 hello 2011-10-23T17:31:49 You create a new one the first time you run I believe. 2011-10-23T17:32:08 I have a problem with following the tutorial with one of the starter kits 2011-10-23T17:32:25 *** SqrT__ has quit IRC (Quit: Page closed) 2011-10-23T17:32:27 *** EspenG has joined #aichallenge 2011-10-23T17:32:34 there is no "isvisible" method as is described in the tutorials 2011-10-23T17:32:38 *** trashant has quit IRC (Quit: Page closed) 2011-10-23T17:32:41 and it's kinda needed 2011-10-23T17:33:26 Which language dinosoep? 2011-10-23T17:34:58 aichallenge: janzert * rf0fa6e0 / website/profile_games.php : Stop processing if profile_games doesn't find user or submission data - http://git.io/-jqJQg 2011-10-23T17:35:38 d 2011-10-23T17:35:43 the d programming language 2011-10-23T17:36:31 some guy posted a message in the newsgroup saying d was officially supported but I think the implementation is still incomplete? 2011-10-23T17:37:06 Yap, seems like it, if you look at ants.py in the python package you can see what they are doing there and reproduce it fairly easily in D I would image dinosoep 2011-10-23T17:38:14 Oh gosh, having to try and read python :p I'll do it and try to translate everything to d, thank you 2011-10-23T17:38:18 *** onensora has joined #aichallenge 2011-10-23T17:38:49 * AndrewBC dons trollface 2011-10-23T17:38:57 is reading pseudocode hard? :D 2011-10-23T17:39:02 * AndrewBC takes it back off 2011-10-23T17:39:41 * Puj briefly smeels troll hide, and quickly dismisses the illusion 2011-10-23T17:40:03 hurr hurr 2011-10-23T17:40:03 xd it's suposed not to be :) 2011-10-23T17:40:15 *** puddy has joined #aichallenge 2011-10-23T17:40:31 Puj: did you ever see this movie? http://www.imdb.com/title/tt1740707/ 2011-10-23T17:41:05 Nope, but my dad recommended it :P AndrewBC 2011-10-23T17:41:19 hehe, and now there are 2 recommenders 2011-10-23T17:41:32 Right up my alley 2011-10-23T17:42:19 trolls be crazy 2011-10-23T17:42:28 *** dinosoep has quit IRC (Quit: Page closed) 2011-10-23T17:43:01 *** cutoff has joined #aichallenge 2011-10-23T17:43:12 *** than_ has quit IRC (Quit: Page closed) 2011-10-23T17:43:29 ugh 2011-10-23T17:43:41 I've gone from having a PERFECT algorithm that is too slow to run 2011-10-23T17:43:55 *** ztfw has quit IRC (Read error: Connection reset by peer) 2011-10-23T17:43:58 to having the FASTEST algorithm that will never succeed. 2011-10-23T17:44:05 hurr hurr 2011-10-23T17:44:19 *** danielharan has joined #aichallenge 2011-10-23T17:44:21 It's better to lose the game than to be disqualified :) 2011-10-23T17:45:04 I once wrote a sudoku solver which tried to fill in the blank squares by brute force. It never solved a puzzle after running for days and days... :( 2011-10-23T17:45:49 *** Puj has quit IRC (Ping timeout: 240 seconds) 2011-10-23T17:46:01 Craklyn: i once got a solver to pass in a programming competition 2011-10-23T17:46:12 Craklyn: all i did was write a check function, and then i filled the puzzle with random numbers 2011-10-23T17:46:14 If you want to make a sudoku solver, just make 27 sets of integers. 2011-10-23T17:46:26 One for each column, row, and 3x3 box 2011-10-23T17:46:35 Craklyn: time limit was 20seconds, but somehow it guessed it before the time was up 2011-10-23T17:46:46 then, make each square in the 9x9 reference 3 of those sets 2011-10-23T17:46:59 the value of the 9x9 is the only number that can fit into all 3 sets 2011-10-23T17:47:03 amstan: hah 2011-10-23T17:47:04 I agree if you want to do it intelligently that's the way to do it Axoren. :P 2011-10-23T17:48:14 This is a great bit on soduko solvers… http://norvig.com/sudoku.html 2011-10-23T17:48:15 if you actually want to solve sudoku, read norvig's post 2011-10-23T17:48:16 efb 2011-10-23T17:48:24 Right now I'm helping an elderly man with Java 2011-10-23T17:48:31 aw dammit 2011-10-23T17:48:37 the enrollment to that ai class ended 2011-10-23T17:48:38 He also has ADHD 2011-10-23T17:48:41 *** danielharan has quit IRC (Ping timeout: 258 seconds) 2011-10-23T17:48:47 I didn't realize it would close enrollment 2011-10-23T17:48:51 Axoren: it was open for 2 months 2011-10-23T17:48:52 :| stupid procrastination 2011-10-23T17:48:57 AndrewBC: ^ 2011-10-23T17:49:00 >:| 2011-10-23T17:49:06 wait... what? 2011-10-23T17:49:09 bah 2011-10-23T17:49:10 Axoren: wp 2011-10-23T17:49:15 lol 2011-10-23T17:49:20 there's probably opencourseware for ai 2011-10-23T17:49:23 or some other resource 2011-10-23T17:49:31 * AndrewBC hunts 2011-10-23T17:49:51 How do you guys deal with movement? 2011-10-23T17:50:35 I know about left wall and derivatives 2011-10-23T17:50:38 but I don't like getting stuck on islands 2011-10-23T17:51:12 and if I look too far ahead, my stuff takes to long and loses 2011-10-23T17:51:17 *** n4nd0 has joined #aichallenge 2011-10-23T17:51:28 if I don't look far enough ahead, I walk into dead ends and freak out 2011-10-23T17:51:34 hmmm, cool there is an AI class on opencourseware 2011-10-23T17:52:34 I can route N ants to the nearest of M resources on a grid with S cells in O(NS) time 2011-10-23T17:52:39 if you don't want to get stuck in dead ends, maybe you could store your path for N moves, backtrack, and strike a new direction 2011-10-23T17:52:53 *** Naploop has joined #aichallenge 2011-10-23T17:53:48 *** vasily has joined #aichallenge 2011-10-23T17:54:32 So explore everything by bruteforcing the map 2011-10-23T17:54:41 as a tree/web 2011-10-23T17:54:46 AndrewBC, there's also ai-class.com. registration is closed, but the videos are on youtube.com/knowitvideos 2011-10-23T17:55:00 *** sromanov has quit IRC (Quit: Ухожу Ñ Ð¾Ñ‚ ваÑ) 2011-10-23T17:55:35 *** bmh has quit IRC (Quit: bmh) 2011-10-23T17:55:38 cyphase: if I remember correctly that was just a wrapper around a course that needed no enrollment. I remember that though... gotta hunt it down again 2011-10-23T17:56:19 AndrewBC, it's based on it, but i believe the videos are being recorded specifically for this, for whatever that's worth 2011-10-23T17:56:51 oh hm apparently it did have a schedule 2011-10-23T17:56:54 http://www.stanford.edu/class/cs221/online.html 2011-10-23T17:57:19 cyphase: ai-class.com also has quizzes and homeworks 2011-10-23T17:57:23 cyphase: cool 2011-10-23T17:57:33 rmmh, yea, i'm registered :) 2011-10-23T17:57:53 oh that's the wrong TLD 2011-10-23T17:57:58 bah 2011-10-23T17:58:20 oh wait sorry 2011-10-23T17:58:29 I can route N ants to the nearest of M resources on a grid with S cells in O(S) + O(N) time 2011-10-23T17:59:41 gonna get to work on my map broker 2011-10-23T17:59:49 so my ants can move atomically 2011-10-23T17:59:56 yay concurrency 2011-10-23T18:00:25 *** Axoren has quit IRC (Quit: Page closed) 2011-10-23T18:01:15 *** bmh has joined #aichallenge 2011-10-23T18:01:19 *** bmh has joined #aichallenge 2011-10-23T18:01:26 ugh, why are my ants fungible? :( 2011-10-23T18:01:36 *** seth has joined #aichallenge 2011-10-23T18:01:45 god forbid they got unique ids. 2011-10-23T18:02:06 bmh, implement it yourself :P 2011-10-23T18:02:20 cyphase: I'm going to, that's why I'm griping! 2011-10-23T18:02:24 lol 2011-10-23T18:02:32 bmh: the server would have to do id renumbering so it doesn't leak information about total number of ants 2011-10-23T18:02:42 heh 2011-10-23T18:02:46 or use UUIDs I guess? 2011-10-23T18:02:53 idk, it's not that hard to remap your ants 2011-10-23T18:02:57 rmmh: I don't care about other people's ants. I care about my ants. 2011-10-23T18:03:05 it indeed would be nice if each ant had an unique id 2011-10-23T18:03:05 oh, right :) 2011-10-23T18:03:11 provided by the server 2011-10-23T18:03:13 or only show ids to the owner 2011-10-23T18:03:16 who needs ids 2011-10-23T18:03:22 when each ant gets a goroutine 2011-10-23T18:03:26 \o/ 2011-10-23T18:04:04 Sorry, I'm dense. Was that pun on coroutine? Or some reference to the Go programming language? 2011-10-23T18:04:19 go's lightweight threads (coroutines) are called goroutines 2011-10-23T18:04:21 *** pedrosorio has quit IRC (Ping timeout: 265 seconds) 2011-10-23T18:04:22 yeah it's Go's version of a coroutine 2011-10-23T18:04:28 specifically based on CSP 2011-10-23T18:04:29 threads? 2011-10-23T18:04:44 nah, they're multiplexed on top of N threads 2011-10-23T18:05:25 *** olexs1 has quit IRC (Quit: Leaving.) 2011-10-23T18:05:28 I wonder why they didn't give each ant an id 2011-10-23T18:05:53 because they're stingy with numbers 2011-10-23T18:05:57 there are only so many, you know 2011-10-23T18:06:10 just track the ants 2011-10-23T18:06:19 http://www.youtube.com/watch?v=drE5cHe6c3s 2011-10-23T18:06:20 i wish the starter kits wouldn't be so crap 2011-10-23T18:06:30 this "bot" abstraction is just moot 2011-10-23T18:06:41 cyphase: at least I should be able to reuse the "Which ants are my ants?" code to figure out "Which ants do my opponents own?" 2011-10-23T18:06:53 yeah, if the starter kit already did the tracking IDs wouldn't be necessary 2011-10-23T18:06:59 * avdg doesn't believe that the smartest js bot is as smart as his first version (aka not moving) 2011-10-23T18:07:15 http://aichallenge.org/language_profile.php?language=Javascript 2011-10-23T18:07:34 oh noes, poor js 2011-10-23T18:07:48 AndrewBC: NOOOO! He wrote an infinite loop :( 2011-10-23T18:07:59 bmh: :D 2011-10-23T18:08:01 that would be a timeout :p 2011-10-23T18:08:15 AndrewBC: is that Kentish Town...? 2011-10-23T18:08:23 bmh: no idea 2011-10-23T18:08:50 wow, 38 2011-10-23T18:08:51 not bad 2011-10-23T18:08:53 oh, nvm, it was a bit clever, it maybe had a bad start on 1 game 2011-10-23T18:09:12 mine doesn't hunt food :p 2011-10-23T18:09:16 its still random 2011-10-23T18:09:26 but bit better than the default random 2011-10-23T18:09:28 if i tell the server to move an ant, will it update me next time with this change? 2011-10-23T18:09:32 *** ermau has left #aichallenge 2011-10-23T18:09:40 i.e. i go issue_order and then will Ants.map reflect this, when? 2011-10-23T18:09:52 lericson: if the move was legal, yes 2011-10-23T18:10:31 mine just randomizes direction when someone hits a wall :p 2011-10-23T18:10:45 wtf http://ants.fluxid.pl/replay.2757 locutusx @turn 210 starts emiting a long column of ants, and suddenly ants go to different directions 2011-10-23T18:11:24 locutus 2011-10-23T18:11:26 nice name 2011-10-23T18:12:13 i wonder, is anyone with a decent bot using any of the starter code? 2011-10-23T18:12:26 locutusx's bot is crazy :-) 2011-10-23T18:12:38 avdg: It looks like I'm watching space invaders. 2011-10-23T18:13:41 4 hills per bot with some top bots: http://aichallenge.org/visualizer.php?game=4567 2011-10-23T18:13:49 Fluxid: I assume that's your site? Is it based on accoun's code? 2011-10-23T18:14:26 *** ikaros_ has quit IRC (Quit: Ex-Chat) 2011-10-23T18:14:50 I shall have... personalities. 2011-10-23T18:14:51 rmmh: my server, code from https://github.com/Ashod/ants-tcp 2011-10-23T18:15:15 Momo bot is vicious and delicious. 2011-10-23T18:15:44 *** pedrosorio has joined #aichallenge 2011-10-23T18:16:00 i saw on replay that 2 ants kill 1 ant without loosing one of them, it's some rule 2 on 1 or atacked ant just forget to atack?) 2011-10-23T18:16:11 vasily: read the battle resolution page 2011-10-23T18:16:24 *** choas has quit IRC (Ping timeout: 240 seconds) 2011-10-23T18:16:38 *** paulwal has joined #aichallenge 2011-10-23T18:17:01 have the problems with the server been fixed yet? 2011-10-23T18:17:09 paulwal: which problems? 2011-10-23T18:17:38 vasily: 1) server calculates how many enemies are within range for each ant 2) an ant dies if it has more enemies in range than any of its enemies have in range 2011-10-23T18:18:15 I just signed up yesterday, but wasn't there like a 2000 hour wait time between games? 2011-10-23T18:18:18 err, that might be ambiguous 2011-10-23T18:18:41 *** pedrosorio has quit IRC (Client Quit) 2011-10-23T18:18:43 i shall test my current and older version of bot... 2011-10-23T18:18:46 vasily: 1) server calculates how many enemies are within range for each ant 2) an ant dies if any of its enemies have fewer enemies in range than the ant has in range 2011-10-23T18:19:03 *** tr3v has joined #aichallenge 2011-10-23T18:19:04 *** pedrosorio has joined #aichallenge 2011-10-23T18:19:29 *** tr3v is now known as Guest99854 2011-10-23T18:19:46 http://aichallenge.org/visualizer.php?game=2237&user=361 <- that was almost another invader 2011-10-23T18:21:48 i like how your ants march in single file 2011-10-23T18:22:05 thats not my bot :p 2011-10-23T18:22:09 *** Epona has joined #aichallenge 2011-10-23T18:22:15 ah 2011-10-23T18:22:20 *** cirno_the_greate has joined #aichallenge 2011-10-23T18:22:28 is there a python visualizer out there? 2011-10-23T18:22:36 2.7 2011-10-23T18:22:49 that's a seriously dumb bot. 2011-10-23T18:23:02 im not a bot :( 2011-10-23T18:23:10 bmh: it's doing pretty well... 2011-10-23T18:23:16 Epona: Exactly what a bot would say. 2011-10-23T18:23:23 rmmh: oljanx is? 2011-10-23T18:23:26 beep 2011-10-23T18:23:30 boop 2011-10-23T18:23:31 *** lurker_for_life has quit IRC (Quit: Page closed) 2011-10-23T18:23:52 aren't you a godess? 2011-10-23T18:24:00 yes 2011-10-23T18:24:31 *** Harpyon has quit IRC (Quit: Textual IRC Client: http://www.textualapp.com/) 2011-10-23T18:24:44 a female programmer... more rare than a unicorn sighting! 2011-10-23T18:24:52 oh, thought you were talking about locutus 2011-10-23T18:25:12 paulwal: not really. 2011-10-23T18:25:27 you must have avoided college 2011-10-23T18:25:30 bmh, reallu? 2011-10-23T18:25:47 I did indeed 2011-10-23T18:26:10 In grad school the department was nowhere near parity, but females were by no measure rare. 2011-10-23T18:26:16 we have a female? 2011-10-23T18:26:22 ffs. 2011-10-23T18:26:32 :) 2011-10-23T18:26:44 *** prgrmr-xtreem has joined #aichallenge 2011-10-23T18:26:46 i'm off to pick up my topology exam :| 2011-10-23T18:26:53 good luck 2011-10-23T18:27:03 bmh, yeah but that's in grad school. how many unicorns were there? 2011-10-23T18:27:12 *** ikaros has quit IRC (Quit: Ex-Chat) 2011-10-23T18:27:12 why does the python starter bot use time.clock() to determine the time so far instead of time.time()? 2011-10-23T18:27:30 paulwal: alive or dead? 2011-10-23T18:27:31 more accurate on windows, cyphase 2011-10-23T18:27:40 cyphase: difference between clock and cpu time maybe? 2011-10-23T18:27:43 Generally we just killed the unicorns for their horns or meat. 2011-10-23T18:27:44 hey dudes, i tried running my program but it doesn't work: "python ants.java" 2011-10-23T18:27:47 c'mon man. unicorns are immortal 2011-10-23T18:27:53 lericson, it's an entirely different function in linux and windows 2011-10-23T18:28:05 then how do you explain all the dead unicorns? 2011-10-23T18:28:32 Epona, wrt. python visualizer: if you find one, tell me about it too :P 2011-10-23T18:28:34 they just need a tear drop from a mermaid to restore them to health 2011-10-23T18:28:49 ill try to work on it 2011-10-23T18:28:52 *** EspenG has left #aichallenge ("Leaving") 2011-10-23T18:28:54 http://www.youtube.com/watch?v=2wRaFXBDn0g 2011-10-23T18:28:55 i don't really want to hack java code 2011-10-23T18:29:20 but adding the ability to visualize debug information woulb be cool 2011-10-23T18:29:26 Alexer: what? 2011-10-23T18:29:31 s/woulb/would/ 2011-10-23T18:29:35 heh, i didn't catch the reference. nice 2011-10-23T18:29:39 on linux time.clock() seems to be cpu time. it's nowhere near actual time passed. that would be time.time() 2011-10-23T18:29:55 *** Guest99854 has quit IRC (Remote host closed the connection) 2011-10-23T18:30:02 guys, i work at google, i know what i'm doing 2011-10-23T18:30:03 lericson, time.clock(), it does entirely different things on linux and windows 2011-10-23T18:30:05 you have to use a turing machine 2011-10-23T18:30:27 can i have a bite of your free google lunch? 2011-10-23T18:30:28 s/turing/universal turing/g 2011-10-23T18:30:40 i can share some guice 2011-10-23T18:30:55 Alexer: yup 2011-10-23T18:31:02 *** Blkt has quit IRC (Quit: cya) 2011-10-23T18:31:17 Alexer: on windows, clock() is better for taking deltas of time than time.time() is 2011-10-23T18:31:22 google? Is that sort like yahoo? 2011-10-23T18:31:24 that is accurate deltas of time 2011-10-23T18:31:25 *** Naploop has quit IRC (Ping timeout: 265 seconds) 2011-10-23T18:31:25 *** kozlovsky has quit IRC (Ping timeout: 265 seconds) 2011-10-23T18:31:57 otzi, google it 2011-10-23T18:32:01 but the servers are linux, so time.time(), right? 2011-10-23T18:33:06 lericson, so it should rather be (time.clock() if sys.platform == 'win32' else time.time()) or something 2011-10-23T18:33:23 *** iron4o has left #aichallenge 2011-10-23T18:34:02 are the servers still backed up by thousands of hours? 2011-10-23T18:34:09 Alexer: sounds about right, i don't know why they made it like that. i remember a discussion on the ML about it 2011-10-23T18:35:09 i'm going to have to remake this ants.py 2011-10-23T18:35:13 it sucks so hard it isn't even funny :( 2011-10-23T18:35:29 How much memory MyBot can eat? 2011-10-23T18:35:50 1 gig i think, according to the forums 2011-10-23T18:36:13 *** n4nd0 has quit IRC (Quit: Leaving) 2011-10-23T18:36:48 lericson, i was most annoyed by the lack of tracking of unseen area :P 2011-10-23T18:36:48 Any reason why the haskell antworld only have Me, Enemy1, Enemy2, Enemy3? Did someone get lazy? 2011-10-23T18:36:51 http://aichallenge.org/visualizer.php?game=4709&user=31 <- check the walls that are created in this game 2011-10-23T18:37:07 big fight, but almost no shots :p 2011-10-23T18:37:25 avdg: GarySWest is an assassin! 2011-10-23T18:38:04 oh, and the fact that you dont inherit from Ants, passing it around as an argument was annoying 2011-10-23T18:38:19 bmh: oljanx is running the starter bot in that match above 2011-10-23T18:38:27 which is why it's so dumb :) 2011-10-23T18:38:30 how is everyone communicating between ants 2011-10-23T18:38:44 lericson, was it those, or do you have something to add to the list? :P 2011-10-23T18:38:46 Epona: Why communicate between ants? You're omnipotent! 2011-10-23T18:38:58 ants don't scale :p 2011-10-23T18:39:02 Epona: I'm organizing my ants in a hierarchy. 2011-10-23T18:39:03 they only help 2011-10-23T18:39:13 oh interesting 2011-10-23T18:39:19 are there bishops and rooks 2011-10-23T18:39:27 only kings. 2011-10-23T18:39:39 but queens are faster 2011-10-23T18:39:46 *** prgrmr-xtreem has quit IRC (Quit: Page closed) 2011-10-23T18:40:13 the problem with allowing ants to communicate is that they begin to procrastinate 2011-10-23T18:40:32 otzi: And then they waste company time. 2011-10-23T18:40:35 Epona, communicating between ants? wasn't the point to make them run around like headless chickens? at least that's what my bot does. ..damn 2011-10-23T18:40:49 i see the colony as one mind and the individual ants are just tools at my disposal. 2011-10-23T18:41:09 no communication necessary 2011-10-23T18:41:09 I see some people make them reproduce too much 2011-10-23T18:41:10 mwuahahaha 2011-10-23T18:41:18 and they lose that way 2011-10-23T18:41:28 how so? 2011-10-23T18:41:29 Epona: reproduce too much? 2011-10-23T18:41:39 in terms of communication or breeding? 2011-10-23T18:41:44 * avdg likes to put the map in a tree 2011-10-23T18:41:45 they'll be hanging around the water cooler while you're queen is slaughtered 2011-10-23T18:41:46 breeding 2011-10-23T18:41:54 your* 2011-10-23T18:42:05 *** SharkMonkey has joined #aichallenge 2011-10-23T18:42:07 I tried to make my ants reproduce, but they ended up killing each other. I think ants are lousy at sex. 2011-10-23T18:42:19 :D 2011-10-23T18:42:21 Epona: 'eh? Why? 2011-10-23T18:42:29 if I had the map saved 2011-10-23T18:42:59 it was on a small map 2011-10-23T18:43:23 otzi, xDD 2011-10-23T18:43:30 *** tillj has quit IRC (Ping timeout: 265 seconds) 2011-10-23T18:44:02 Epona, i can't see how having more ants could ever make you perform worse 2011-10-23T18:44:17 im still learning the way it works here 2011-10-23T18:44:21 Alexer: if your program is dumb and you time out? 2011-10-23T18:44:34 bmh, ...touche :D 2011-10-23T18:45:26 its silly how many people lock up 2011-10-23T18:45:31 http://ai-contest.com/visualizer.php?game=4560 2011-10-23T18:45:50 :D 2011-10-23T18:46:26 Epona: those bots are hurting Britney :( 2011-10-23T18:46:40 For the harvest? 2011-10-23T18:46:57 not enough zerg rush 2011-10-23T18:47:18 how long does it usually take for a new starter package to be accepted? 2011-10-23T18:47:50 *** jtamer_ has quit IRC (Quit: Leaving...) 2011-10-23T18:49:13 protip: a timeout goroutine and a select loop make timing out practically impossible 2011-10-23T18:49:36 you can even be greedy and use as much time as you can every turn, consistently, and run optional things after the necessary things 2011-10-23T18:49:43 greed be awesome 2011-10-23T18:50:02 *** jtamer has joined #aichallenge 2011-10-23T18:50:34 is the turn timeout always the same or does it vary? 2011-10-23T18:50:43 it looks like those bots were very simplistic anyway. i doubt they were doing anything too heavy. they also all timed out at about the same time. i suspect that the server is to blame for those timeouts 2011-10-23T18:51:04 paulwal: it is allowed to vary, but i think right now they are the same 2011-10-23T18:51:17 Epona, you don't *have* to timeout on large number of bots: http://ants.fluxid.pl/replay.2702 2011-10-23T18:51:18 which is usualy 1 sec? 2011-10-23T18:51:22 yeah 2011-10-23T18:51:25 right now, at least 2011-10-23T18:51:32 ok 2011-10-23T18:51:33 you get turntime as input 2011-10-23T18:51:34 it's a parameter because it might change in the future 2011-10-23T18:52:25 yea i see the parameter. i haven't actually uploaded and tested a bot yet. apparently there was like a 2000 hour wait time yesterday 2011-10-23T18:53:00 this one is amazing 2011-10-23T18:53:00 *** pvarga has left #aichallenge 2011-10-23T18:53:27 if you time out it's just a problem with the programming, like with my bot here: http://ants.fluxid.pl/replay.2754 2011-10-23T18:53:49 on tcp it can be network fault too 2011-10-23T18:53:50 because i was lazy and i only check the timeout between long calculations, not during them 2011-10-23T18:53:57 Fluxid, ah, true that 2011-10-23T18:54:10 and turn time is longer on tcpserver 2011-10-23T18:54:21 can has packet loss plx 2011-10-23T18:54:44 does anyone working with python have an example they can show me? 2011-10-23T18:55:07 don't worry, I promise i'm no competition at all 2011-10-23T18:55:07 Fluxid, in my case turn time being longer shouldn't affect timeouts that much, since i try to abort if i have less than 100ms time remaining :P 2011-10-23T18:55:11 *** jtamer has quit IRC (Read error: Connection reset by peer) 2011-10-23T18:55:25 code sharing for things not very alike to the starter bots is frowned upon 2011-10-23T18:55:32 AndrewBC: Why...? 2011-10-23T18:55:45 I dunno, I tried and got rebuked 2011-10-23T18:55:54 By whom? 2011-10-23T18:55:57 bmh: the server is already flooded with starter bots, they don't want it flooded with slightly-more-sophisticated bots 2011-10-23T18:56:01 the contest admins 2011-10-23T18:56:06 bmh, it reads on the site 2011-10-23T18:56:12 *** irchs has quit IRC (Quit: irchs) 2011-10-23T18:56:16 rmmh: That's a reasonable rationale. 2011-10-23T18:56:33 i think they just don't want a working decent bot posted and then copy & pasted by 1000 noobies 2011-10-23T18:56:45 bmh, sharing ideas and discussing is okay, but sharing ready-made code (aside from starter packs) is discouraged 2011-10-23T18:56:46 bmh: basically they want rank to be entirely dependent on things you personally write 2011-10-23T18:56:48 I don't plan on doing that 2011-10-23T18:57:17 im just not used to seeing python using multiple files 2011-10-23T18:57:21 my python bot processes 50 ants in 150ms 2011-10-23T18:57:22 it's odd to me 2011-10-23T18:58:12 if there tons of identical bots it will just clog up the server and degrade the contest. No one wants to see 90% of the top 100 bots be exact copies of eachother 2011-10-23T18:58:21 bmh, yeah basically what they said 2011-10-23T18:58:27 darn afkness distracted me 2011-10-23T18:58:40 Epona: how much python have you written? 2011-10-23T18:58:40 paulwal: They should hash the bots and kill dupes. 2011-10-23T18:58:50 I write a lot of things in pygame 2011-10-23T18:58:56 and just use one file 2011-10-23T18:59:04 *** HaraKiri has quit IRC () 2011-10-23T18:59:09 global variables eh 2011-10-23T18:59:21 I also work with django 2011-10-23T18:59:22 *** seth has quit IRC (Quit: Page closed) 2011-10-23T18:59:22 they do or they should? I agree they should. Of course some could just make a small change. And then there's all the duplicate starter bots 2011-10-23T18:59:35 that uses multiple files, but isn't really as relevant 2011-10-23T18:59:42 a depressing number of new games are of 8 bots all going straight up and then crashing into each other 2011-10-23T18:59:56 if a small change doesn't result in any hash change, then you've got a bad hash for these purposes 2011-10-23T19:00:20 hash the bosh outputs :) 2011-10-23T19:00:24 bot outputs 2011-10-23T19:00:32 ah yea 2011-10-23T19:00:34 yeah, you could precompute the input of each bot 2011-10-23T19:00:43 and then precompute a map existing solely of starter bots 2011-10-23T19:00:52 er, a game output for each map 2011-10-23T19:01:02 a1k0n: hey 2011-10-23T19:01:07 then it would be no strain at all for people to upload starter bots 2011-10-23T19:01:08 there would still inevitably be duplicate bots with adjustable constants 2011-10-23T19:01:13 bmh, while i admit i had the same idea, it goes against the idea that it should be easy to test everything is working and start seeing results 2011-10-23T19:01:16 amstan: sup 2011-10-23T19:01:19 a1k0n: are you planning on compeeting again? 2011-10-23T19:01:21 a1k0n: starter bot deduplication, clever 2011-10-23T19:01:33 bmh, step 1: upload bot; step 2: site tells you "no can do, duplicate bot" 2011-10-23T19:02:04 amstan: i am. don't know how serious i'll be about it but i have some awesome ideas that are sure to be terrible once implemented 2011-10-23T19:02:10 Alexer: it would be nice if the site directed you to play on an official tcp server 2011-10-23T19:02:11 *** antong has joined #aichallenge 2011-10-23T19:02:16 a1k0n: nice nice 2011-10-23T19:02:17 instead of submitting starter bots to the ladder 2011-10-23T19:02:26 hi guys, anyone using java starter package? 2011-10-23T19:02:26 of course you could special-case it and show the matches from the identical bot, but in that case at some point something would break 2011-10-23T19:03:06 *** Katiska has joined #aichallenge 2011-10-23T19:03:13 rmmh, yeah. i suggested creating an official tcp server to the organizers, but they're too busy currently with the normal server :P 2011-10-23T19:03:19 hehe 2011-10-23T19:03:23 how about to get into the top tier rankings, your code must differ by a certain percentage from all other bots in the top X ranks 2011-10-23T19:03:27 detecting file duplicates is a solved problem by SCM tools 2011-10-23T19:03:29 Fluxid: how much load is the tcp server? 2011-10-23T19:03:45 just do that for every file used in a bot, it'll be hashes 2011-10-23T19:03:55 rmmh, but the problem isn't really starter bots per se. i mean, if the starter bots weren't uploaded, there'd still be enough users to clog up the servers 2011-10-23T19:04:07 all match = duplicate, any mismatch = early bailout because not dupe 2011-10-23T19:04:14 rmmh, it just happened sooner because everyone submitted a starter bot before their real one 2011-10-23T19:04:30 Alexer: yeah. The starter guide said "Submit the stock bot" oops :) 2011-10-23T19:04:34 rmmh: 1.6% of CPU 2011-10-23T19:04:34 yeah maybe the 5-minute quickstart shouldn't be "download this package" then "upload this package" 2011-10-23T19:04:45 Fluxid: so no need to optimize, cool 2011-10-23T19:04:46 rmmh: out of all 8 cores 2011-10-23T19:04:55 AndrewBC, but there could and would be dupes that simply change a constant or add a comment to avoid creating a duplicate hash 2011-10-23T19:05:10 paulwal: those result in hash changes 2011-10-23T19:05:11 *** cybsy has quit IRC (Ping timeout: 244 seconds) 2011-10-23T19:05:23 otherwise git wouldn't know anything when I change a constant or add a comment 2011-10-23T19:05:26 :P 2011-10-23T19:05:38 since it does, the hash must have changed 2011-10-23T19:05:48 it might even use two different hashing algorithms, I forget 2011-10-23T19:06:03 AndrewBC, right...so they wouldn't be seen as duplicates, right? 2011-10-23T19:06:14 bmh, it's easy to perceive that as the problem, but it isn't a problem. the perceived problem results from the fact that most haven't updated their first starter-bot-submission to a real bot yet 2011-10-23T19:06:37 Alexer: I can only write code so fast :) 2011-10-23T19:06:41 paulwal: oh I see what you're saying 2011-10-23T19:06:56 you'd want to detect users who are intentionally trying to bypass the restriction 2011-10-23T19:06:58 AndrewBC, i'm saying it would still suck if there were hundreds of almost duplicate bots with just a simple change of some constants 2011-10-23T19:07:11 hi anyone had problems with java starter package ? 2011-10-23T19:07:11 bmh, and i'm saying it isn't a problem :) people just perceive it as a problem 2011-10-23T19:07:18 Depends how many people upload unedited bots... 2011-10-23T19:07:27 I guess flag bots with extremely similar file structure 2011-10-23T19:07:30 Antong: What's the problem? Mine mostly works, and I use java 2011-10-23T19:07:36 I suppose I should upload an unedited one 2011-10-23T19:08:05 oh well, I don't think it's a problem 2011-10-23T19:08:17 yea 2011-10-23T19:08:20 Craklyn: when I do, java -jar MyBot.jar it says could not find the main class 2011-10-23T19:08:20 the starter bots are a problem only if a big percentage of the users are only going to upload a starter bot and then leave it at that 2011-10-23T19:08:22 if it becomes one, they can always just pore over the submission source 2011-10-23T19:08:28 and disqualify if need be 2011-10-23T19:08:36 Antong: Haha, I had that problem too 2011-10-23T19:08:44 Craklyn: but if i just do: java MyBot it works 2011-10-23T19:08:48 the quickstart literally tells you to just submit a starter package, so it shouldn't be hard to just use hashes of the archives 2011-10-23T19:08:56 I spent about 6 hours today dealing with it and othe rproblems. What are you running your java in? Windows command line, cygwin, linux? 2011-10-23T19:09:15 Craklyn: it's ubuntu 2011-10-23T19:09:23 Alexer: the matchmaking doesn't matchmake inactive bots 2011-10-23T19:09:27 runing from command line 2011-10-23T19:09:38 that could probably save a lot of time 2011-10-23T19:09:41 I never could get it to work in cygwin (linux). When I run jara -jar MyBot.jar in windows command line it works. 2011-10-23T19:09:45 *** Novarg has left #aichallenge 2011-10-23T19:09:49 Sorry, I don't have a solution for you on that :( 2011-10-23T19:10:03 AndrewBC: interesting workaround with the truncated 2011-10-23T19:10:06 AndrewBC, inactive defined by what? it only gives N games to every bot before ignoring it completely? 2011-10-23T19:10:14 oh well ok, will try in windows then 2011-10-23T19:10:17 AndrewBC: you'll have to keep in contact with janzert for it 2011-10-23T19:10:21 amstan: thanks, it's kind of a hack :D 2011-10-23T19:10:38 AndrewBC: it is? it makes the output bigger for the first test, right? 2011-10-23T19:10:39 "not a real solution" et al, but good enough 2011-10-23T19:10:54 amstan: yeah 2011-10-23T19:11:07 *** bmh has quit IRC (Quit: bmh) 2011-10-23T19:11:08 but it's conceivable that it might not be enough for someone else, so it might have to be changed again 2011-10-23T19:11:18 Craklyn: ok thanks mate, see you 2011-10-23T19:11:35 but oh well, I don't think you guys would appreciate me putting the entire error log for every test into the db ;x 2011-10-23T19:11:55 *** foRei has quit IRC (Read error: Connection reset by peer) 2011-10-23T19:12:33 amstan: did janzert get a notification for it? He'll probably see it soon if he hasn't already 2011-10-23T19:13:15 I'm not in too big a hurry, gives me time to make algorithmic changes to the bot 2011-10-23T19:13:21 what type of game strategy is everyone leaning towards? 2011-10-23T19:13:44 I'm leaning towards emergent behavior from individualized ants 2011-10-23T19:13:57 probably won't win unless I'm awesome, but it'll be fun :) 2011-10-23T19:13:58 i'm trying to get my 11-year old brother to start programming 2011-10-23T19:14:03 what do you mean by that? 2011-10-23T19:14:16 AndrewBC: he might get a notification for it, you can wait a little and see what happens 2011-10-23T19:14:28 AndrewBC: in the meantime, what language are you having problems with? 2011-10-23T19:14:30 I mean I'm giving every ant the capability to make their own decisions based on behavior sets and tweaks within the behaviors 2011-10-23T19:14:45 cool 2011-10-23T19:14:48 *** djr_ has quit IRC (Ping timeout: 240 seconds) 2011-10-23T19:15:08 amstan: just a go runtime error during the server test. My client side test doesn't though :/ it's strange 2011-10-23T19:15:10 so no communication or master plan, just each ant deciding based on it's visible squares? 2011-10-23T19:15:24 *** daniloresende has quit IRC (Ping timeout: 265 seconds) 2011-10-23T19:15:26 AndrewBC: the server runtime might be different 2011-10-23T19:15:33 AndrewBC: what language? 2011-10-23T19:15:38 cyphase: there will be a little bit of communication or master plan, but only like setting ratios of behaviors 2011-10-23T19:15:41 amstan: Go 2011-10-23T19:15:58 * avdg wishes everything could be coded in 1 loc 2011-10-23T19:16:13 *** orfest has joined #aichallenge 2011-10-23T19:16:18 hi all 2011-10-23T19:16:22 I'm just coding too slow :p 2011-10-23T19:16:23 rmmh: hurr hurr. At first I thought it was because my tools were old, but I updated them and it doesn't seem that was it 2011-10-23T19:16:28 it can be. don't know how readable it'll be though 2011-10-23T19:16:46 ; is your friend 2011-10-23T19:16:47 meh readable locs ;-) 2011-10-23T19:16:48 AndrewBC: did you version match? 2011-10-23T19:17:07 amstan: version of what? 2011-10-23T19:17:11 of go 2011-10-23T19:17:12 *** Slater has joined #aichallenge 2011-10-23T19:17:14 "sudo readMyBrain" 2011-10-23T19:17:25 avdg, one loc: curl -O http://aichallenge.org/starter_packages/c_starter_package.zip 2011-10-23T19:17:25 ah, should be r60.something something right? 2011-10-23T19:17:29 lemme spit it out 2011-10-23T19:17:36 AndrewBC: not sure 2011-10-23T19:17:56 6g version release.r60.1 9498 2011-10-23T19:18:19 Could you give any insight on my problem? Whenever I submit my bot (which is written from scratch and works perfectly on a local machine, even test_bot.sh works ok), being submitted to the server after reading "player_seed 42" reads only empty strings? Instead of "ready" and then "turn 1" it receives just nothing. 2011-10-23T19:18:21 *** garlicnation has quit IRC (Quit: garlicnation) 2011-10-23T19:18:31 I don't remember where I saw the version the server uses 2011-10-23T19:19:26 ah 6g version release.r60.1 9753+ 2011-10-23T19:19:39 that 9753 can't be right though... 2011-10-23T19:19:50 why not 2011-10-23T19:20:31 the go repository is up to 10152 2011-10-23T19:21:03 *** anty_m has joined #aichallenge 2011-10-23T19:21:21 my local copy is 60.3 9516 so 9753 sounds wrong 2011-10-23T19:21:43 yeah, 9512 is 60.2 2011-10-23T19:21:54 *** dlila has quit IRC (Remote host closed the connection) 2011-10-23T19:22:15 if you do `hg tags` in your go src repo you'll see what I mean 2011-10-23T19:23:55 aichallenge: Alexandru Stan * r2a0d0c7 / ants/dist/starter_bots/java/Makefile : Really fixed java Makefile. Added jar to be a proper target in the makefile - http://git.io/vkF9Cw 2011-10-23T19:24:16 *** kinetic has joined #aichallenge 2011-10-23T19:24:35 seeing if I get a runtime error with go @ 9573 2011-10-23T19:24:48 6g version weekly.2011-08-17 9573 2011-10-23T19:24:51 is its version 2011-10-23T19:25:43 well I don't, but I'm still not ruling out that that might be why 2011-10-23T19:26:33 *** roflmao has left #aichallenge 2011-10-23T19:27:02 I'll just print the server version on my next upload at runtime 2011-10-23T19:27:04 hurr hurr 2011-10-23T19:28:51 *** dd has joined #aichallenge 2011-10-23T19:33:41 *** dd has quit IRC (Client Quit) 2011-10-23T19:34:02 *** kinetic has quit IRC (Quit: Page closed) 2011-10-23T19:36:11 *** GreenTea has quit IRC (Ping timeout: 265 seconds) 2011-10-23T19:36:30 bleh, I won because the other bots died :/ 2011-10-23T19:37:17 for a prior challenge, the server gave more game opportunities to players whose skill level was less confidently known... is that the case this time? 2011-10-23T19:38:38 *** letmetellyou has joined #aichallenge 2011-10-23T19:38:45 *** Larose has quit IRC (Quit: Leaving.) 2011-10-23T19:39:03 when a game starts, do you immediately know all the locations of the walls of the game? 2011-10-23T19:39:21 letmetellyou, no 2011-10-23T19:39:43 *** onensora has quit IRC () 2011-10-23T19:42:35 another question: in the recent games display, my outcome is always 1st of , even when I was eliminated. And actually, in the opponents column, they're all listed as 1st. How should I interpret this? 2011-10-23T19:42:42 tl;dr - How do I know if I won? 2011-10-23T19:43:14 by looking at the score bar? 2011-10-23T19:43:27 http://aichallenge.org/specification.php#Scoring 2011-10-23T19:43:55 so I have to go into the viewer for each game to see the result? 2011-10-23T19:44:00 *** vasily has quit IRC (Quit: Page closed) 2011-10-23T19:44:14 not really, you can check the status 2011-10-23T19:44:20 *** bitbrit has quit IRC (Quit: bitbrit) 2011-10-23T19:44:44 (which still requires you to browse the lists) 2011-10-23T19:44:51 hmm, okay, maybe I just got paired up against bots that were just as dumb as mine (the default bot) 2011-10-23T19:44:57 thanks 2011-10-23T19:45:05 hehe 2011-10-23T19:45:11 there are way too many of them 2011-10-23T19:46:46 so how do I move my ant north 2011-10-23T19:46:54 just by putting, "n"? 2011-10-23T19:47:12 o (row) (col) (dir) 2011-10-23T19:47:40 http://aichallenge.org/specification.php#Bot-Output 2011-10-23T19:49:27 so you specify which bot 2011-10-23T19:49:32 and then a direction 2011-10-23T19:49:38 s/bot/ant/ 2011-10-23T19:49:41 yeah 2011-10-23T19:50:01 cool 2011-10-23T19:50:49 after making changes "for better" my bot got -10 in skill :( 2011-10-23T19:51:11 :p 2011-10-23T19:51:22 mine bot is still under heavy construction 2011-10-23T19:51:46 (which got nuked few times) 2011-10-23T19:52:10 i have problems with not colliding own ants in tight groups :| 2011-10-23T19:52:36 * avdg with implementing the algorithm 2011-10-23T19:52:46 its too much typing 2011-10-23T19:55:38 You have to see enemy hill locations before you're told of them, yes? 2011-10-23T19:55:46 Do you see captured hills? 2011-10-23T19:55:54 nope 2011-10-23T19:55:56 yes, no 2011-10-23T19:56:10 unless you compute it yourself 2011-10-23T19:56:24 >:) 2011-10-23T19:56:33 *** mjard has left #aichallenge 2011-10-23T19:56:43 so capping gives you points and hides info from enemies 2011-10-23T19:56:44 nice 2011-10-23T19:57:23 depends on what you mean with hiding information :p 2011-10-23T19:57:53 if you capture a hill quickly, the other bots might not even know there ever as a hill there 2011-10-23T19:58:04 yeah 2011-10-23T20:02:49 would a map that is surrounded by water violate the symmetry constraint? Probably, unless it's 2 or 4 players? 2011-10-23T20:03:18 * avdg likes a map that separates players 2011-10-23T20:03:32 *** jtamer has joined #aichallenge 2011-10-23T20:03:47 *** levis501 has joined #aichallenge 2011-10-23T20:03:59 i wonder if it's better in python to read in line by line, or in chunks 2011-10-23T20:04:05 i'm sort of leaning towards chunks 2011-10-23T20:04:13 cyphase: what's a chunk? 2011-10-23T20:04:42 *** nack has quit IRC (Quit: Page closed) 2011-10-23T20:04:45 http://aichallenge.org/specification.php#Map-Format @map generator 2011-10-23T20:04:55 amstan, i just mean all the input available at the time 2011-10-23T20:04:59 you'll be better off reading bytes that are a multiple of disk read sizes 2011-10-23T20:05:10 AndrewBC, but it's not from the disk 2011-10-23T20:05:24 cyphase: why is this important? 2011-10-23T20:05:29 *** pvarga_ has joined #aichallenge 2011-10-23T20:05:41 AndrewBC, i guess i didn't specify. i'm talking about the bot input 2011-10-23T20:05:50 mm 2011-10-23T20:05:51 amstan, performance possibilities 2011-10-23T20:06:19 i/o optimizations will not give you enough of a speedup 2011-10-23T20:06:22 http://aichallenge.org/visualizer.php?game=5103 <- I like this ending 2011-10-23T20:06:37 most time is consumed during processing, not i/o 2011-10-23T20:06:42 *** ApocalypticAnt has joined #aichallenge 2011-10-23T20:06:54 hi everyone 2011-10-23T20:07:07 amstan, i know, but why not if it actually helps 2011-10-23T20:07:28 well, it's all about time spent on your optimizations 2011-10-23T20:07:33 why not start with the most costly thing 2011-10-23T20:07:36 *** pvarga_ is now known as pvarga 2011-10-23T20:07:37 *** blackace227 has joined #aichallenge 2011-10-23T20:07:52 well, thing is i have nothing to optimize yet 2011-10-23T20:07:55 :P 2011-10-23T20:07:57 profile! 2011-10-23T20:08:00 cyphase: i wouldn't do it, unless it was 5% of your turn or more 2011-10-23T20:08:08 it just came to mind while going over ants.py 2011-10-23T20:08:10 cyphase: yes, get some times in there and profile it 2011-10-23T20:08:15 Is it wrong this is my homework for class? 2011-10-23T20:08:20 nono, i mena i have nothing to profile 2011-10-23T20:08:22 I was wondering if I could get some help running play_game inside tutorial.cmd 2011-10-23T20:08:25 no code besides the tutorial bot 2011-10-23T20:08:27 blackace227: hmm? 2011-10-23T20:08:28 cyphase: get to work! :P 2011-10-23T20:08:31 lol 2011-10-23T20:08:33 yes, i am 2011-10-23T20:08:40 * AndrewBC cracks the whip 2011-10-23T20:08:46 yes'm! 2011-10-23T20:08:50 oh lawdy 2011-10-23T20:09:33 *** blackace227 has quit IRC (Client Quit) 2011-10-23T20:09:39 *** ants9238483 has quit IRC (Quit: Page closed) 2011-10-23T20:09:50 When i change the --map_file option in the tutorial.cmd file it only runs for one turn 2011-10-23T20:10:08 but when i use the example map it runs fine 2011-10-23T20:11:10 it is failiing at the python line: if self.map_similar(hill_aims[0][0], loc, aim, enemy_map[0]) is None: KeyError 0 2011-10-23T20:14:10 *** antsnub has joined #aichallenge 2011-10-23T20:14:17 ApocolypticAnt: I had similar errors when I gave the wrong path to a map 2011-10-23T20:14:24 hey everyone, will the c++ gettimeofday function work on the official servers? 2011-10-23T20:14:52 ah ok ty anty_m 2011-10-23T20:15:20 np 2011-10-23T20:15:46 when i specify a completely wrong map name it comes up with a different error 2011-10-23T20:16:00 *** mceier has quit IRC (Quit: leaving) 2011-10-23T20:16:25 hmm 2011-10-23T20:16:28 this is the option and path i am setting --map_file tools/maps/maze/maze_02p_01.map 2011-10-23T20:16:42 avdg: awesome! 2011-10-23T20:16:45 but getting error 2011-10-23T20:17:19 oh, wait... I was getting that error when I was messing with the food options maybe? 2011-10-23T20:17:34 avdg: lol 2011-10-23T20:17:47 yeah i did change it to 4,8 2011-10-23T20:17:52 as opposed to none 2011-10-23T20:18:35 i did change it to none and the visualiser came up this time 2011-10-23T20:18:39 but still for one turn 2011-10-23T20:21:02 can someone output a command on windows to run a game (play_game) on one of the formail maps 2011-10-23T20:21:09 formal* 2011-10-23T20:22:17 bweh, default bots aren't useless http://aichallenge.org/visualizer.php?game=5123 2011-10-23T20:22:21 *** swinejihad has joined #aichallenge 2011-10-23T20:22:54 that's awesome 2011-10-23T20:23:43 avdg: i wonder why it wasn't terminated with "hills not being razed" 2011-10-23T20:23:48 *** ApocalypticAnt has quit IRC (Quit: Page closed) 2011-10-23T20:23:59 maybe because they didn't saw it? 2011-10-23T20:24:16 um, it's automatic? 2011-10-23T20:24:16 dunno how the algo works though 2011-10-23T20:24:26 i saw similar games with defaultbots 2011-10-23T20:24:35 *** replore_ has joined #aichallenge 2011-10-23T20:24:39 and those were terminated quickly... 2011-10-23T20:24:41 bah. uploaded my bot, have been waiting for an hour for my first game, only 4.5 hours remaining. :-/ 2011-10-23T20:24:46 it's really embarassing that my bot doesn't know it should defend its hill 2011-10-23T20:25:10 just put an ant on top of it :p 2011-10-23T20:25:17 brilliant idea 2011-10-23T20:25:17 cool, still around 30rd place 2011-10-23T20:27:46 *** az__ has joined #aichallenge 2011-10-23T20:28:13 a 5-hour evaluation loop is going to force a ton of local testing and ~1 upload/night. bummer. 2011-10-23T20:29:07 *** az__ has quit IRC (Client Quit) 2011-10-23T20:30:04 someone should make a genetic algorithm to make antbots... 2011-10-23T20:30:46 sounds like fun 2011-10-23T20:31:56 or machine learning 2011-10-23T20:32:40 *** McLeopold has joined #aichallenge 2011-10-23T20:33:23 *** murilo has joined #aichallenge 2011-10-23T20:33:36 someone using the cpp starter kit? 2011-10-23T20:34:02 i am 2011-10-23T20:34:04 kinda 2011-10-23T20:34:30 i'm.. mostly ripping it to shreds 2011-10-23T20:34:30 I want to know how I can know if there is my ant in a Square 2011-10-23T20:34:50 ant == 0 2011-10-23T20:34:54 hmm 2011-10-23T20:34:59 if there's no ant it's -1 2011-10-23T20:35:02 I thought was > 0 2011-10-23T20:35:07 >0 are enemy 2011-10-23T20:35:08 nope, you're player 0 2011-10-23T20:35:08 what are you using? 2011-10-23T20:35:17 thank you 2011-10-23T20:35:29 nvrmind 2011-10-23T20:35:38 *** noam has quit IRC (Ping timeout: 258 seconds) 2011-10-23T20:35:47 that is a naive method for hill defense. leave an ant on the hill to prevent spawning. then when someone sends ants to attack, you keep respawning an ant as they get destroyed 2011-10-23T20:35:53 ah, >0 are enemy in the protocol, dunno about cpp 2011-10-23T20:35:55 of course, if they come at you with any kind of formation.. 2011-10-23T20:36:23 how to keep equal distribution of ants on the map? 2011-10-23T20:36:36 *** bobby_ has joined #aichallenge 2011-10-23T20:36:40 any algorithms to do that? 2011-10-23T20:36:44 I have to map a position to a ant number 2011-10-23T20:36:49 knapsack algorithm :P 2011-10-23T20:36:57 i think if you can define "equal distribution of ants" you can find such an algorithm 2011-10-23T20:37:44 could use flocking/swarming algorithms to keep a certain distance, etc. 2011-10-23T20:37:51 Fluxid: i though the seed passed to you on setup is your own player id? 2011-10-23T20:38:07 awishformore: no, that's arbitrary afaik 2011-10-23T20:38:24 so why would you even need the seed? 2011-10-23T20:38:36 to make your random bot deterministic when testing 2011-10-23T20:38:41 a1k0n: i want to keep as much of map visible as possible. to do that i want distances between ants to be roughly equal... 2011-10-23T20:39:05 so your own ants always get id 0? 2011-10-23T20:39:12 Fluxid: so i'd start with a greedy algorithm to maximize the number of seen squares 2011-10-23T20:39:15 awishformore: yep 2011-10-23T20:39:22 (i am working on that very thing right now) 2011-10-23T20:39:36 i mean, also on the server right? 2011-10-23T20:39:44 not just talking test suites here 2011-10-23T20:40:02 the documentation is really lacking ^^ 2011-10-23T20:40:06 but hmm, i guess the euclidean distance between ants is equivalent 2011-10-23T20:40:32 a1k0n: maybe i'll do the same algorithm i use to do map discovery... to make ants move where is something invisible 2011-10-23T20:40:51 *** ashoka has joined #aichallenge 2011-10-23T20:41:06 but floodfilling from edges is... not great 2011-10-23T20:41:26 *** bobby_ has quit IRC (Ping timeout: 265 seconds) 2011-10-23T20:41:44 *** pipeep has quit IRC (Ping timeout: 276 seconds) 2011-10-23T20:41:46 *** csharp__ has joined #aichallenge 2011-10-23T20:42:13 Anyone else on thats using C# and has it working with the tools? 2011-10-23T20:43:15 that's what i'm doing now -- distance to unseen squares, which means all the ants decide to go the same way 2011-10-23T20:43:45 did anyone using C/C++ have issues with timing out because he did too many/slow computations? 2011-10-23T20:43:57 i'm sure someone did 2011-10-23T20:44:14 *** tester has joined #aichallenge 2011-10-23T20:44:22 setitimer is your friend there... at least get _some_ moves made rather than timing out 2011-10-23T20:44:37 Is it ok if I ask some basic questions about underlying bot setups? 2011-10-23T20:44:45 ask away 2011-10-23T20:44:52 well 2011-10-23T20:44:54 *** Redgis has quit IRC (Ping timeout: 244 seconds) 2011-10-23T20:44:58 Does anyone know what spawnradius2 is? 2011-10-23T20:44:58 I have nil AI experience and wanted to run a few ideas around 2011-10-23T20:45:01 *** djr_ has joined #aichallenge 2011-10-23T20:45:21 i probably won't time out once i'm done ;) 2011-10-23T20:45:24 tester, how close food has to be for you to eat it (and spawn a new ant) 2011-10-23T20:45:27 tester: it's unused i think 2011-10-23T20:45:36 it's always 1 2011-10-23T20:45:42 I'm guessing no, but is trying to keep track of each "ant" as an individual agent a good or even feasible idea? 2011-10-23T20:45:43 tester: holdover from when ants used to spawn from food, rather than from a hill 2011-10-23T20:45:44 *** smartypanda has joined #aichallenge 2011-10-23T20:45:45 but i'm wondering if it's worth it to look at ways to trade off optimal for speed 2011-10-23T20:45:58 i think it is maximum distance from hill ants can spawn 2011-10-23T20:46:26 tester: spawnradius2 is the square of the distance that an ant must be from food in order to eat it 2011-10-23T20:46:35 the name is wrong for historical reasons :\ 2011-10-23T20:46:36 swinejihad: it's a start. i treat my ants independently currently except for doing dependency resolution when they want to move on top of each other 2011-10-23T20:46:49 Fluxid: ^^ 2011-10-23T20:47:12 jmcarthur: ah! 2011-10-23T20:47:30 Thanks for all the replies! Are the values for viewradius and attackradius guaranteed to be constant over time? 2011-10-23T20:47:38 yes 2011-10-23T20:47:40 the name comes from a time when there were no hills and food simply turned into an ant when you were in range 2011-10-23T20:47:41 *** Dominater has joined #aichallenge 2011-10-23T20:47:43 Over the competition period? 2011-10-23T20:47:48 no 2011-10-23T20:47:55 tester: guaranteed to be a constant for the duration of a single game 2011-10-23T20:47:56 over game period - yes 2011-10-23T20:47:57 ;) 2011-10-23T20:48:21 *** csharp__ has quit IRC (Quit: Page closed) 2011-10-23T20:48:33 I mean it in the sense of say maybe wrapping an "ant" in an object, so I could say give one ant tile an objective to go somewhere over many turns 2011-10-23T20:49:03 does anyone use D* instead of A* with moving ants as dynamic obstacles? 2011-10-23T20:49:08 Has anyone seen a range of values for viewradius and attackradius? Is viewradius always greaterthan attackradius? By how much? What are the current values? 2011-10-23T20:49:46 i think some people are doing that. i'm not. 2011-10-23T20:49:58 swinejihad: i think that many OO styles are bad for performance, but not necessarily bad design aside from that 2011-10-23T20:50:13 i always replan from scratch every frame. shrug. 2011-10-23T20:50:17 *** skyramp has joined #aichallenge 2011-10-23T20:50:18 swinejihad: you might have a hard time maintaining that abstraction though, since you would have to reidentify your ants on each turn 2011-10-23T20:50:39 I decided against it because of the complexity of trying to figure out where the ant is/whether it got killed every turn 2011-10-23T20:50:42 jmcarthur: that's not all too hard 2011-10-23T20:50:46 jmcarthur: exactly my thoughts 2011-10-23T20:50:57 awishformore: it might be too hard if you have to ask 2011-10-23T20:52:31 wooo, only 3 hours out of 5.5 remaining until my bot's first server game. 2011-10-23T20:52:55 the servers are running slow again. maybe some good games coming up. 2011-10-23T20:53:16 still so much work left on my bot 2011-10-23T20:53:19 can't wait to put it on ;) 2011-10-23T20:53:43 a 10-player map going to 1000 turns with 1 second per turn each... could take almost 3 hours to run the game 2011-10-23T20:54:11 if they weren't well parallelized 2011-10-23T20:54:22 does it run turns in parallel? 2011-10-23T20:54:29 are these multi-cpu instances? 2011-10-23T20:54:36 2 core, i think 2011-10-23T20:54:38 i think it tries to run one bot per core at a time 2011-10-23T20:54:43 oh ok 2011-10-23T20:54:48 i would guess that most bots usually respond way faster than 1 second though 2011-10-23T20:54:54 i guess that cuts it down to 1.5 hrs ;) 2011-10-23T20:55:09 you usually want to respond quicker 2011-10-23T20:55:30 if people start doing alpha-beta searches on battle clusters, though, it could get slow 2011-10-23T20:55:38 and tremendously interesting to watch 2011-10-23T20:55:44 * jmcarthur looks back and forth suspiciously 2011-10-23T20:55:49 heh. 2011-10-23T20:56:09 did i say that out loud? 2011-10-23T20:56:25 *** letmetellyou has quit IRC (Ping timeout: 265 seconds) 2011-10-23T20:56:54 *** djr_ has quit IRC (Read error: Connection reset by peer) 2011-10-23T20:57:18 *** djr_ has joined #aichallenge 2011-10-23T20:57:37 i don't think that would be the best approach though :P 2011-10-23T20:58:02 the best approach is a hard thing to prove 2011-10-23T21:00:26 i don't think it's a bad idea 2011-10-23T21:01:47 i kind of think all these attempts by various people to incorporate A* into their AI are misguided. either that, or i'm missing something. 2011-10-23T21:01:59 i concur 2011-10-23T21:02:11 but i don't begrudge them for using it 2011-10-23T21:02:42 *** paulwal has quit IRC (Quit: Page closed) 2011-10-23T21:03:07 for all i know there's some awesome tactic that uses A* to good effect 2011-10-23T21:03:17 jmcarthur, why misguided? 2011-10-23T21:03:20 my bot does a few BFS's and that's it. takes 20ms/frame on a big map for any number of ants. it's a stupid bot, though. 2011-10-23T21:03:39 cyphase: why good does it serve? 2011-10-23T21:03:43 *what good 2011-10-23T21:03:54 *** djr__ has joined #aichallenge 2011-10-23T21:04:53 jmcarthur, not sure, i haven't gotten to that yet, but presumably pathfinding. i was just wondering what your critique was 2011-10-23T21:05:34 you can't do A* until you have a source and destination. you can instead compute global information on the map in one pass, like distance to the nearest food for instance 2011-10-23T21:05:35 cyphase: A* works very well when you have clear goals, but it's not very often in this game that your goals are so clear. 2011-10-23T21:05:36 *** djr_ has quit IRC (Ping timeout: 240 seconds) 2011-10-23T21:05:49 *** Islacrusez has joined #aichallenge 2011-10-23T21:06:50 it's scary that my sample bot is going up in ranks... o.O 2011-10-23T21:06:58 a1k0n: there are extensions to a* to be multi-agent and multi destination 2011-10-23T21:07:19 not very well published, unfortunately... 2011-10-23T21:07:42 an intelligent algorithm (a* or breadth-first) is still useful for short term path-finding 2011-10-23T21:07:47 aichallenge: Alexandru Stan * r32bc2f3 / (ants/dist/starter_bots/Makefile ants/dist/tools/Makefile): Included VERSION files in tools and starter packages. - http://git.io/hgLxEA 2011-10-23T21:07:55 without them you end up with bots like mine :( 2011-10-23T21:09:10 a simple BFS works very well for local pathfinding. A* works well, too. it's just not going to give much additional benefit 2011-10-23T21:09:17 hey swine, remember me? 2011-10-23T21:09:18 a1k0n, distance to nearest food in one pass? o.O 2011-10-23T21:09:31 or in fact, anyone I was bugging last night? xD 2011-10-23T21:10:06 check that VERSION file out now: http://aichallenge.org/starter_packages/cpp_starter_package.zip 2011-10-23T21:10:09 *** FlyinFish has joined #aichallenge 2011-10-23T21:10:16 all autogenerated, for all starter packages, and the tools zip 2011-10-23T21:10:22 Games per minute per server 0.4 0.2 0.5 ........ :'( 2011-10-23T21:10:54 *** skyramp has quit IRC (Quit: Page closed) 2011-10-23T21:11:00 woah, a pull request for an Ada starter pack? 2011-10-23T21:11:00 anty_m: say thanks there's more games than users right now, before it was lower 2011-10-23T21:11:04 jmcarthur: yeah 2011-10-23T21:11:23 @amstan wow, ok. I'll stop complaining 2011-10-23T21:11:24 anty_m: I'm sorry Dave, err anty_m; I cannot 'amstan'. 2011-10-23T21:11:24 i use haskell and i say that's hardcore :P 2011-10-23T21:11:56 *** swinejihad__ has joined #aichallenge 2011-10-23T21:12:03 whoops, back 2011-10-23T21:12:04 is there any particular reason my test_bot won't even run? o.O 2011-10-23T21:12:08 ayo swine 2011-10-23T21:12:22 *** swinejihad has quit IRC (Ping timeout: 265 seconds) 2011-10-23T21:12:32 *** Moduspwned has joined #aichallenge 2011-10-23T21:12:35 the actual bot files doesn't do anything when run 2011-10-23T21:12:50 do we have an ada compiler? 2011-10-23T21:13:03 there's a python command that starts it all up 2011-10-23T21:13:05 Zannick: the pull request also changed the setup script to include it 2011-10-23T21:13:07 *** Migi32 has quit IRC (Quit: bye) 2011-10-23T21:13:11 amstan: oh, nice 2011-10-23T21:13:11 i note that the vast majority of that patch is devoted to numerous copies of the apache license... 2011-10-23T21:13:12 which would beeeee? 2011-10-23T21:13:17 Zannick: after i told him 2011-10-23T21:13:33 Alexer: sure. start with all the food on the map, add the locations to list1. dist=0. foreach item in list1 { mark distance=dist. add all neighbors that haven't been visited already to list2} list1=list2; clear list2; repeat until list1 is empty. 2011-10-23T21:13:36 just go through the tutorial on the website 2011-10-23T21:13:46 I want to see how my new bot does, but it has been 5 hours since I submitted it and still not games. 2011-10-23T21:13:57 the tutorial has been helpful up to the point where I actually downloaded the files 2011-10-23T21:13:59 Alexer: visits each square on the map once. it's basically unit-cost dijkstra. 2011-10-23T21:14:00 The rest of you need to have your bots take less time ;-) 2011-10-23T21:14:06 or.. breadth-first-search. 2011-10-23T21:14:46 a1k0n, ah, we just had different definitions for "one pass" :P 2011-10-23T21:15:20 Alexer: well, heh. it's O(N) for N squares on the map. 2011-10-23T21:15:32 a1k0n, though i do it by bfs:ing from each food to nearest ant one food at a time 2011-10-23T21:15:33 I tried calling my bot in play_one_game.sh and it always returns as crashed 2011-10-23T21:15:34 but yeah, not like top to bottom 2011-10-23T21:16:05 Alexer: it's probably a good idea if you're allocating ants to food, cuz otherwise you have to redo it to keep all your ants from going to the same food 2011-10-23T21:16:07 a1k0n, it shouldn't be that much slower, since usually you eat food quite fast and nearest ant is always inside viewradius 2011-10-23T21:16:58 *** nick has joined #aichallenge 2011-10-23T21:17:02 you don't know food locations until you're within viewradius of food ...right? 2011-10-23T21:17:13 anty_m, yep 2011-10-23T21:17:15 correct 2011-10-23T21:17:24 *** nick is now known as Guest76893 2011-10-23T21:17:27 *** TerryVB has joined #aichallenge 2011-10-23T21:17:27 but it could be on the other side of water in a maze, so you might lose sight of it temporarily 2011-10-23T21:18:00 Hey guys, is there a known bug with the coffeescript starter kit? 2011-10-23T21:18:18 Is there some issue that has caused the number of games/min to drop? Or is it just the number of people who submitted? 2011-10-23T21:18:56 *** jtamer has quit IRC (Quit: Leaving...) 2011-10-23T21:19:07 Islacrusez: I'm just not very familiar with OSX, nor am I particularly good at computers compared to everyone else here 2011-10-23T21:19:09 *** smartypanda has quit IRC (Ping timeout: 265 seconds) 2011-10-23T21:19:14 *** Rinum has joined #aichallenge 2011-10-23T21:19:43 *** pvarga has quit IRC (Ping timeout: 248 seconds) 2011-10-23T21:20:06 Moduspwned: I'd guess both of those times the fact that the bots needed to play these games take a long time to play each turn 2011-10-23T21:20:37 swinejihad__: they should decrease the time allowed then. 2011-10-23T21:20:56 when does this contest end? I'm currently busy with school and probably won't be able to submit any code for a few weeks 2011-10-23T21:21:04 Dec 18, I think? 2011-10-23T21:21:10 yes, ish 2011-10-23T21:21:55 *** FlyinFish is now known as analyst74 2011-10-23T21:22:20 anty_m: lol my b-day 2011-10-23T21:22:45 nice 2011-10-23T21:23:01 Moduspwned: it would be interesting to see how people's bots react to that but the contest needs to balance rewarding efficient code and playing a lot of people versus rewarding optimal AI strategies 2011-10-23T21:23:05 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T21:23:31 Moduspwned: you can just connect to one of the tcp servers to see how your new bot performs 2011-10-23T21:23:40 Moduspwned: in anycase I think it's one second per player per turn 2011-10-23T21:23:47 if I was organizing it, I would make a seperate total, and per-turn limit. like.. limit of 1second per turn, and you get 10seconds plus .1seconds extra each turn 2011-10-23T21:23:55 analyst74: oh? where do I find out about one of those? I didnt see anything on the website. 2011-10-23T21:24:16 swinejihad__: one second per turn?! holy cow, I was thinking 10 times less than that 2011-10-23T21:24:51 I was thinking like 0.1 seconds per turn, I guess I can be far more ambitious 2011-10-23T21:24:54 need to let java run its GC 2011-10-23T21:25:02 http://aichallenge.org/forums/viewtopic.php?f=25&t=1505 2011-10-23T21:25:10 Moduspwned: I'm just going by hearsay 2011-10-23T21:25:25 analyst74: cool, I will check that out. 2011-10-23T21:25:26 analyst74: bookmarked, thank you. 2011-10-23T21:25:43 np, the more the merrier 2011-10-23T21:26:03 Moduspwned, if you want to get your bot to play games against others, put it on http://ants.fluxid.pl/howto and/or http://ash.webfactional.com/howto ;) 2011-10-23T21:26:14 ya, I was going to be really sad if I had to wait ~6-8 hours for any feedback every time I upgrade 2011-10-23T21:26:15 *** k333es has joined #aichallenge 2011-10-23T21:26:24 Alexer: thanks! 2011-10-23T21:26:56 this is very dangerous 2011-10-23T21:27:03 awesome 2011-10-23T21:27:07 giving others your code 2011-10-23T21:27:14 cyphase, what is? 2011-10-23T21:27:25 bots on the internet. at some point someones bot is gonna get loose and become skynet 2011-10-23T21:27:34 you don't give out your code, you run your bot on your machine over a tcp socket 2011-10-23T21:27:51 oh, duh... X( 2011-10-23T21:28:08 cyphase: you know, that gets funnier each time we hear it... 2011-10-23T21:28:11 getting sleepy, I guess. :-/ will have to check my rankings tomorrow 2011-10-23T21:28:23 *** Epona has quit IRC (Quit: Leaving) 2011-10-23T21:28:45 Islacrusez, i actually lol'ed at that 2011-10-23T21:29:08 *** mleise has quit IRC (Ping timeout: 245 seconds) 2011-10-23T21:30:00 *** Olap has quit IRC (Ping timeout: 256 seconds) 2011-10-23T21:30:09 *** avers_ has joined #aichallenge 2011-10-23T21:30:48 *** Olap has joined #aichallenge 2011-10-23T21:31:50 *** mleise has joined #aichallenge 2011-10-23T21:31:52 Hmm, I think I must be doing something wrong on the tcp client 2011-10-23T21:32:10 *** cyphase has quit IRC (Quit: http://www.cyphase.com/) 2011-10-23T21:33:01 *** icefox has joined #aichallenge 2011-10-23T21:33:21 *** xathis_ has quit IRC () 2011-10-23T21:33:46 *** cyphase has joined #aichallenge 2011-10-23T21:35:34 *** peregrinator has quit IRC (Ping timeout: 265 seconds) 2011-10-23T21:36:09 Oh..... maybe I should have an ant playing defense 2011-10-23T21:36:40 my thinking was the best defense is a good offence 2011-10-23T21:36:42 Brilliant, ok one ant will always play D, so it will take 2 ants for me to lose 2011-10-23T21:36:45 turns out to not actually work 2011-10-23T21:36:57 ya, I found all the games I lost my hill was to a single ant 2011-10-23T21:36:59 should be easy to fix 2011-10-23T21:37:31 my last game was embarassing that way: http://aichallenge.org/visualizer.php?game=5139&user=432 2011-10-23T21:37:43 orange is wiped out, but he has three ants wandering around the map who eventually catch me with my pants down 2011-10-23T21:38:50 also my ants mostly killed each other 2011-10-23T21:39:12 sometimes my bot seems to do nothing on the tcp, like just first ant doesnt move. I lagged out of the game or something? 2011-10-23T21:39:56 odd, not sure 2011-10-23T21:40:04 are you playing on ants.fluxid.pl? 2011-10-23T21:40:34 I am running on both, maybe thats the problem 2011-10-23T21:40:49 should be fine 2011-10-23T21:41:52 *** skyramp has joined #aichallenge 2011-10-23T21:41:59 *** Rinum has quit IRC (Quit: Page closed) 2011-10-23T21:42:16 wait.. does this map have hills that are walled in?? http://ash.webfactional.com/replay.859 2011-10-23T21:42:58 Also, what is the logic of which hill an ant is generated on when a food is captured? 2011-10-23T21:43:00 looks like it 2011-10-23T21:43:47 I am pretty happy, I have absoultly no search algorithm implimented at all. No bfs, no a*, I wonder how far I can get before caving and writing one 2011-10-23T21:43:48 *** swinejihad__ has quit IRC (Ping timeout: 266 seconds) 2011-10-23T21:44:55 *** danielharan has joined #aichallenge 2011-10-23T21:45:45 yeah those maps are gone from the official server but the tcp servers still use them 2011-10-23T21:46:18 i think it's round-robin on hills that are unblocked 2011-10-23T21:46:51 whichever hill has least-recently had an ant move off it 2011-10-23T21:46:59 skyramp is a bit defensive with its ants 2011-10-23T21:47:40 Moduspawned: I just crashed on ants.fluxid.pl, you might have been affected :p 2011-10-23T21:48:51 and holy cow, a1k0n, you got some interesting combat AI 2011-10-23T21:49:13 i do? 2011-10-23T21:49:23 my bot is utterly unaware of enemy ants 2011-10-23T21:49:47 what it has is an emergent property of the stupid way it goes about collecting food and/or exploring 2011-10-23T21:50:24 ya, your ants are like the vikings "we don't collect food, we just burn villages!" 2011-10-23T21:50:25 ok, the sample bot always crashes on me, any ideas? 2011-10-23T21:50:37 oh yeah, they like smashing anthills too 2011-10-23T21:50:47 but they ignore enemy combatants in the process 2011-10-23T21:51:27 interesting group behaviour though 2011-10-23T21:51:29 Moduspwned: i removed those from ants.fluxid.pl 2011-10-23T21:51:44 yeah it makes defacto squads. i thought it was kinda neat. 2011-10-23T21:51:56 a1k0n: which is your bot? 2011-10-23T21:52:03 http://ants.fluxid.pl/player/a1k0n 2011-10-23T21:52:09 my ants mostly operating independently, efficient gatherers, not quite good at fighting 2011-10-23T21:52:27 ah, didn't see on games list 2011-10-23T21:52:28 yeah i'm about to throw my bot away for the most part 2011-10-23T21:52:38 it's torn apart right now so i haven't run it in a while 2011-10-23T21:52:40 Isacrusez, how is your bot crashing? 2011-10-23T21:53:41 end result reads "crashed" 2011-10-23T21:53:52 I turned off my code that dealt with other ants for now. I couldn't get it to work right yet. 2011-10-23T21:54:12 turn 0 bot 0 crashed Traceback (most recent call last): File "./MyBot.py", line 54, in Ants.run(MyBot()) AttributeError: class Ants has no attribute 'run' 2011-10-23T21:54:43 My favorite thing I have right now is that each ant litterally does no thinking, It only looks at the 4 sqaures next to it. I wish this was score weighted by how many cpu cycles you used. 2011-10-23T21:55:19 surely that defeats the object of artificial "intelligence"? 2011-10-23T21:55:30 def do_turn(ants): pass 2011-10-23T21:56:08 "/usr/bin/yes go" 2011-10-23T21:56:48 Next game should be within 238.4 minutes. Oo 2011-10-23T21:56:57 Isacrusez, are you using bots from the python2 package? 2011-10-23T21:57:10 this means that *my* next game starts within 238.4 minutes? 2011-10-23T21:57:20 I believe so, is there a quick way to tell? 2011-10-23T21:57:31 how are you running it? 2011-10-23T21:58:31 via terminal, modified one of the play_one scripts 2011-10-23T21:58:51 have you tried to run it without modification? 2011-10-23T22:00:21 the script will run fine, can even pick one of the other sample_bots, but the starter bot crashes if I try to run that 2011-10-23T22:00:28 made no modifications to the starter bot 2011-10-23T22:00:37 I have no idea how to begin making my ants handle combat... Looking ahead with a Min/Max algorithm is obviously impossibly time consuming. Any obvious approaches that I don't know about? :P 2011-10-23T22:00:42 Same problem with the Coffee-script starter bot 2011-10-23T22:00:56 a1k0n: i expect a good bot from you :P 2011-10-23T22:01:18 I hope to see crazy formations, flying 'V' and such from people 2011-10-23T22:01:38 hehe, i really want to make multiplayer game of life 2011-10-23T22:01:54 antimatroid: we'll see 2011-10-23T22:01:57 we came up with rules for it but decided it was too hard to watch and complicated for an actual contest 2011-10-23T22:02:27 well, we have rules for 2 player, more players is more complicated because you can have a "draw" from birthing a location 2011-10-23T22:02:49 *** anty_m has quit IRC (Quit: Page closed) 2011-10-23T22:03:42 I just watched a game and was super impressed with my bot, then realized I was watching the wrong color 2011-10-23T22:03:52 lol 2011-10-23T22:03:57 haha 2011-10-23T22:04:00 burny: you're back :D 2011-10-23T22:04:21 i expect a good bot from you within days 2011-10-23T22:04:24 skyramp implemented combat. ouch. 2011-10-23T22:04:35 I proably wont' start till december 2011-10-23T22:04:59 a1k0n: ya skyramp is scrazy, so far on the webfactional site he is the only bot mine has faced 2011-10-23T22:05:01 did I miss any of these contests since last year? 2011-10-23T22:05:15 burny: only been the tron then the planets 2011-10-23T22:05:18 I randomly decided to look a couple days ago, and seems like I barly cought the start of this one 2011-10-23T22:05:28 yeah, it started two days ago 2011-10-23T22:05:38 i'm still not really sure how to do combat, one idea is to work out the maximum number of enemies that could possibly be within the battle radius of each location at the end of the turn 2011-10-23T22:05:38 a1k0n, ya, skyramp is raping my ants on tcp server :( 2011-10-23T22:05:41 burny: It just started, but some people had acess to it early as a beta, I think they are the one dominating right now 2011-10-23T22:06:00 yea 2011-10-23T22:06:03 then try and move my ants as close as possible to enemies while they wont die 2011-10-23T22:06:05 i'm withdrawing my crappy bot until i get some defense 2011-10-23T22:06:20 defense, or offense 2011-10-23T22:06:30 a1k0n: what do you think about markov chains for tracking down runaway enmy ants? 2011-10-23T22:06:36 first priority is not getting razed by errant wandering ants 2011-10-23T22:06:36 I think I am just going to teach mine to be cowards 2011-10-23T22:06:37 i'm really keen to try that out 2011-10-23T22:07:02 * Islacrusez goes off to fetch a cup of tea since he can't even get a starter bot to run unmodified 2011-10-23T22:07:03 antimatroid: like a probability envelope where the bot could be? 2011-10-23T22:07:15 i've been thinking along those lines 2011-10-23T22:07:16 I think you can rank fairly high by not losing any hill instead of razing others :) 2011-10-23T22:07:25 a1k0n, from the tcp servers? please don't. if people take their bots offline, other people have nothing to play against :( defeats the purpose 2011-10-23T22:07:37 just leave it running, pretty please? ;) 2011-10-23T22:07:40 sigh, ok 2011-10-23T22:07:41 yeah, if you see an ant the probability of an enemy being there is 1, then when enemies go out of range just assume they travel in each possible direction with equal probability and work out the transition probabilities 2011-10-23T22:08:03 yeah, except for the part you can see so it's a little skewed 2011-10-23T22:08:16 then if you have ants to spare and you could rank the locations by the probability of an ant being there and go to those places, or weight them based on path length as well 2011-10-23T22:08:26 *** Acorn has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:08:31 a1k0n, i haven't done anything to my bot in 2 days, but i just run it there so that other people always have at least *something* to play against :D 2011-10-23T22:08:37 you can also try to divine their epicenter and hence their hill 2011-10-23T22:08:45 yeah i know, i haven't fully worked it out yet, i'm not actually doing anything until i've handed my thesis in and finished exams 2011-10-23T22:08:45 and you can also infer symmetry to find those 2011-10-23T22:09:03 and maybe give more weight to areas that are more important, e.g. areas closer to your hill 2011-10-23T22:09:04 i've done my best to make it hard to work out the type of symmetry for maps 2011-10-23T22:09:05 *** otzi has quit IRC (Ping timeout: 260 seconds) 2011-10-23T22:09:18 antimatroid: autocorrelation will get it 2011-10-23T22:09:19 a1k0n: yeah symmetry inference is a nice topic 2011-10-23T22:09:21 i'll even translate the map if that'll make it harder 2011-10-23T22:09:22 i should work and sleep, instead i just keep loitering around this channel and doing nothing productive :P 2011-10-23T22:09:28 once you find any hill, you know exactly where all hills are 2011-10-23T22:09:31 except for the part where it's reflected, that's slightly harder but not much 2011-10-23T22:09:46 you can probably find the symmetry in n log n with FFTs 2011-10-23T22:10:01 hmmm, well people are welcome to do it 2011-10-23T22:10:03 or n*(small number of candidates) otherwise 2011-10-23T22:10:14 FFT? that is applicable there? 2011-10-23T22:10:15 i'm a big proponent of if it's possible it's fair game, but i'll make it as hard as possible :P 2011-10-23T22:10:21 a1k0n: oh.. I like that idea. time to fourier some ants 2011-10-23T22:10:30 well not the ants, but the water 2011-10-23T22:10:32 mleise: who knows, sounds fun though 2011-10-23T22:10:39 mleise: convolution theorem 2011-10-23T22:10:46 i gotta go, my kid is whining to play trains 2011-10-23T22:10:49 *** otzi has joined #aichallenge 2011-10-23T22:10:55 i should get back to my exam too 2011-10-23T22:11:00 I thought it would end up being bitwise pattern matching 2011-10-23T22:11:06 *** jtamer has joined #aichallenge 2011-10-23T22:11:10 is analyzing ht eopponents ant moving patterns, to determine which opponent it is, bad? 2011-10-23T22:11:23 Moduspwned, fourier some ants? http://xkcd.com/26/ :) 2011-10-23T22:11:25 adminfreenas 2011-10-23T22:11:27 hehe, I like to play with the trains too, it will become a programmer later 2011-10-23T22:11:30 oops 2011-10-23T22:11:31 burny: probably way more effort than it's worth imo 2011-10-23T22:12:20 i wonder if seeing another ant hill does tell you all information about a tessellation symmetric map? 2011-10-23T22:12:27 *** csyncope has left #aichallenge 2011-10-23T22:12:27 i don't think it necessarily does 2011-10-23T22:12:37 it'll get you a sub tessellation 2011-10-23T22:12:41 at least 2011-10-23T22:12:47 it does 2011-10-23T22:12:55 are you sure? 2011-10-23T22:12:57 i don't think it does 2011-10-23T22:13:09 unless hills are symetric, for a single player also 2011-10-23T22:13:18 ie. the water compared to the hill 2011-10-23T22:13:28 *** dgroft has joined #aichallenge 2011-10-23T22:13:34 but even then.. a bit of exploring solves that 2011-10-23T22:13:36 *** cyphase has quit IRC (Read error: Connection reset by peer) 2011-10-23T22:13:43 say there are 8 players, you might find a second hill such that the tesselation you get is for 4 players 2011-10-23T22:13:48 you might not get all the symmetry 2011-10-23T22:14:00 is the mirroring only done for two players? 2011-10-23T22:14:08 a1k0n: 2, 4 or 8 2011-10-23T22:14:14 hmm 2011-10-23T22:14:14 *** Dominater has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:14:20 there are 5 types of 2 player mirroring maps, 3 for 4 player and 1 for 8 2011-10-23T22:14:21 if it's for only 4.. then the missing 4, are half way inbetween 2011-10-23T22:14:27 and then there's also tesselation symmetry 2011-10-23T22:14:41 burny: sure, but you don't know unless you explore more 2011-10-23T22:14:49 *** cyphase has joined #aichallenge 2011-10-23T22:14:54 my point is seeing another hill doesn't necessarily tell you the number of players 2011-10-23T22:15:00 you can probably infer where all ant hills could be 2011-10-23T22:15:18 but they may be destroyed before you see them too 2011-10-23T22:15:23 yeah you pretty much have to try matching up the water with various candidate symmetries 2011-10-23T22:15:24 ah, I'm not aware of what info's you're given, besides site info 2011-10-23T22:15:36 you aren't told the number of players 2011-10-23T22:15:47 map size? 2011-10-23T22:15:51 yes 2011-10-23T22:15:59 that gives you some information about the number of players 2011-10-23T22:16:03 ya 2011-10-23T22:16:08 er? aren't you told the number of players? 2011-10-23T22:16:11 if we used lcm(players we're going to use) then it's not 2011-10-23T22:16:15 is that only at the end of the game? 2011-10-23T22:16:17 a1k0n: nope :) 2011-10-23T22:16:19 yep 2011-10-23T22:16:20 doohhh 2011-10-23T22:16:29 I think I give up =[ 2011-10-23T22:16:35 well, it's still possible 2011-10-23T22:16:38 Islacrusez: don't say that, what's not working? 2011-10-23T22:16:53 but my bot is not in a state to take advantage of that yet so it's an interesting subproblem for future exploration 2011-10-23T22:17:01 a1k0n: the dimensions give you no info on the number of players if they're divisible by every possible number of players 2011-10-23T22:17:06 antimatroid: i noticed that 2011-10-23T22:17:13 starter bot crashes 2011-10-23T22:17:17 you can check the GCDs 2011-10-23T22:17:23 yep 2011-10-23T22:17:23 *** lrh9 has joined #aichallenge 2011-10-23T22:17:35 it's sad when you can't even get the starting package to run without modification 2011-10-23T22:17:48 Islacrusez: which language? 2011-10-23T22:17:55 *** dungeonduke has joined #aichallenge 2011-10-23T22:17:56 python 2011-10-23T22:18:04 use c++ :P 2011-10-23T22:18:10 ... 2011-10-23T22:18:12 I use python3, and it works fine 2011-10-23T22:18:25 can't say for python2 package though 2011-10-23T22:18:33 analyst74, it worked fine 2011-10-23T22:18:34 Last I checked there was a bug (lol) in ants.py Ants.vision. 2011-10-23T22:18:48 *** jtamer has quit IRC (Quit: Leaving...) 2011-10-23T22:19:11 Islacrusez, you've done something to the ants.py i think 2011-10-23T22:19:20 Islacrusez, Ants should have the run method.. 2011-10-23T22:19:22 this game is a lot more interesting than planets was 2011-10-23T22:19:26 It's easy to fix though. Just change the variable ants.loc to something. (I forget what.) 2011-10-23T22:19:34 burny: in which ways? 2011-10-23T22:19:44 in complexity, for ai 2011-10-23T22:19:47 i think planet wars offered a really awesome problem, but ants has way more depth wrt ai ideas to run with 2011-10-23T22:20:01 last game, was nearly trivial to get first palce 2011-10-23T22:20:08 haha 2011-10-23T22:20:25 just cause you made the top 100 in a week doesn't mean bocsimaco's effort was trivial :P 2011-10-23T22:20:37 shit, i was #1 for the first couple days of planetwars 2011-10-23T22:20:45 yeah but you won tron :P 2011-10-23T22:20:45 ended up, like, >#200 2011-10-23T22:20:52 the trivial solution I was stating, a coupel days after I started.. is what bocs said his solution was 2011-10-23T22:21:25 a1k0n: i meant they started a week before the end of the contest and finished top 100 2011-10-23T22:21:49 burny: implementation was still hard, it was by no means trivial imo 2011-10-23T22:22:02 crap, now I've really buggered the files 2011-10-23T22:22:07 Islacrusez, if you download a fresh python starter package you'll see that the Ants class in ants.py has a run method (and i just checked that to make sure there isn't some new bug) 2011-10-23T22:22:09 time to redownload 2011-10-23T22:22:11 yes, implementation wasn't trivial 2011-10-23T22:22:23 *** xar0l_ has joined #aichallenge 2011-10-23T22:22:26 *** k333es_ has joined #aichallenge 2011-10-23T22:22:37 *** D has joined #aichallenge 2011-10-23T22:22:42 i want to do tron again 2011-10-23T22:22:54 I find the difference between top bots is really small, at least in Tron 2011-10-23T22:23:03 *** D is now known as Guest42467 2011-10-23T22:23:07 i got disqualified in tron :( 2011-10-23T22:23:19 I think this one, will have a MUCH lower frequency, of strong bots losing against weak bots 2011-10-23T22:23:41 http://csclub.uwaterloo.ca/contest/profile_games.php?user_id=2788 after that awesome performance :( 2011-10-23T22:23:54 *** k333es has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:23:58 How did you get disqualifiied? 2011-10-23T22:24:19 it used to be a rule that if you timed out in a game in the final contest you were disqualified 2011-10-23T22:24:23 *** lavalamp has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:24:31 i didn't even know of the barbaric rule until it happened, it was devestating 2011-10-23T22:24:41 Wow 2011-10-23T22:24:52 *** TerryVB has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:24:52 *** Guest76893 has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:24:52 *** Moduspwned has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:24:52 *** antsnub has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:24:54 I haven't even found in the rules what the time constraints are 2011-10-23T22:25:07 1second per playe rpe rmove, I think 2011-10-23T22:25:08 dw, i didn't shut up at the start of planet wars until that was changed 2011-10-23T22:25:14 I've heard a few people say that time is going to be a big issues in this game 2011-10-23T22:25:15 i think it's only 500ms on the server atm 2011-10-23T22:25:21 *** dgroft has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:25:21 *** Slater has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:25:21 *** xar0l has quit IRC (Ping timeout: 265 seconds) 2011-10-23T22:25:28 brb, gonna see if the new set of files runs 2011-10-23T22:25:29 it is a big issue in every game, but yeah it's going to be massive in this one 2011-10-23T22:25:30 ya, it will be.. 2011-10-23T22:25:38 is a lot of actions to consider 2011-10-23T22:25:51 my multi source and multi target a* food collection could take up to 300ms or so during beta 2011-10-23T22:26:02 but i can always tell it to stop collecting after xms 2011-10-23T22:26:14 damned timeouts, forcing me to redesign and optimize my code... I've been at it all day 2011-10-23T22:26:29 ants is more tolerating though, if you time out one game, you don't even necesarrily lose that one 2011-10-23T22:26:31 I'm just trying to grasp the basic concepts like learning what an A* search algorithm is, I'm going to be hopeless when it comes time to optamize 2011-10-23T22:26:32 bah, optimization.. the nerve! :P 2011-10-23T22:26:38 *** icefox has quit IRC (Quit: icefox) 2011-10-23T22:27:04 analyst74: yeah, personally i wanted to set those scores to zero but it's a contest on writing ai strategies not so much prorgamming 2011-10-23T22:27:08 *** icefox has joined #aichallenge 2011-10-23T22:27:29 Guest42467: do you know how to do a simple bfs from a source to a target? 2011-10-23T22:27:52 *** xar0l_ is now known as xar0l 2011-10-23T22:27:57 antimatroid: bfs? 2011-10-23T22:28:04 breadth-first search 2011-10-23T22:28:24 or best first search, depends 2011-10-23T22:28:41 *** ChrisH_ has joined #aichallenge 2011-10-23T22:28:45 antimatroid: not yet, but I've still got plenty of starter material to read through which should help me get to a decent point 2011-10-23T22:28:54 Guest42467: it's easier to add the target to a queue and mark the location as added, then while the source is not found take the front location off the queue and for each of its neighbours, if it's the source you've found a path (move the source into the location you just took off the queue) otherwise if the location has not been added to the queue add it and mark it as added 2011-10-23T22:28:58 *** f4hy has joined #aichallenge 2011-10-23T22:29:00 breadth first search is what i mean too 2011-10-23T22:29:09 that is the easiest useful path finding method 2011-10-23T22:29:32 once you understand and have implemented that then try to understand a*, a* is basically just bfs with a heuristic to speed the search up 2011-10-23T22:29:56 *** skyramp has quit IRC (Quit: Page closed) 2011-10-23T22:30:07 Guest42467, http://www-cs-students.stanford.edu/~amitp/gameprog.html 2011-10-23T22:30:19 *** skyramp has joined #aichallenge 2011-10-23T22:30:21 lots of good stuff there 2011-10-23T22:30:36 WOOOP 2011-10-23T22:30:38 I has bot! 2011-10-23T22:30:43 Islacrusez: :D 2011-10-23T22:30:52 cyphase: cool, I'll check it out 2011-10-23T22:30:59 now to go back to the tutorial and see what to do with it now 2011-10-23T22:31:15 *** avers_ has quit IRC (Remote host closed the connection) 2011-10-23T22:31:16 the tutorial bot cant even beat hunterbot 2011-10-23T22:31:28 no, you have to do that yourself :) 2011-10-23T22:31:36 antimatroid: Went a little over my head. I'll save what you wrote though, and when I get into the actuall programming of it I'll re-read it and see if it makes more sense then 2011-10-23T22:32:01 Guest42467: no, but that really is the easiest path finding method imo 2011-10-23T22:32:02 hunterbot is actually quite good, I didn't beat it until v2 of my bot. :) 2011-10-23T22:32:04 that is useful 2011-10-23T22:32:10 np* 2011-10-23T22:32:32 hunterbot gets stuck the moment it sees food over water 2011-10-23T22:32:41 i had no idea what i was doing at the start of tron, i used one of the tutorial pages to work out how to make a game tree and did that 2011-10-23T22:32:48 I'm still trying to beat hunterbot... 2011-10-23T22:32:51 but I need to know how to edit my bot, and what python code looks like 2011-10-23T22:32:59 How long does food take to disappear? 2011-10-23T22:33:04 it doesn't 2011-10-23T22:33:07 cyphase: quit giving everyone my secret webpage] 2011-10-23T22:33:09 :P 2011-10-23T22:33:21 unless ants belonging to more than one distinct bot are within the spawn radius of it after battle resolution 2011-10-23T22:33:54 Cool, thanks 2011-10-23T22:34:29 yoden, my bad ;) 2011-10-23T22:35:08 it really is the best resource for grid-based a* i've found 2011-10-23T22:35:15 minus my own improvements, which i'll someday blog :X 2011-10-23T22:35:29 i haven't found anywhere with good a*? 2011-10-23T22:35:39 is there anywhere that describes multi target a*? 2011-10-23T22:35:48 antimatroid: that' 2011-10-23T22:35:59 that's one of my improvements, and, no, not to my knowledge :X 2011-10-23T22:36:14 what's this line mean? #!/usr/bin/env sh 2011-10-23T22:36:19 *** ChrisH_ has quit IRC (Quit: Page closed) 2011-10-23T22:36:43 yoden: where abouts? 2011-10-23T22:36:55 oh sorry, misread 2011-10-23T22:37:11 i have a partially written blog post, but i'll never get finished, lol 2011-10-23T22:37:11 yoden: i assume you're using h(x) := min manhatten distance from x to a target? 2011-10-23T22:37:30 Islacrusez, it's *nix related. It's a comment and benign. 2011-10-23T22:37:31 What kind of methods can you use to force your program not to timeout? Are there timer catches that you could add in? 2011-10-23T22:37:45 uhh, basically, plus the cross product optimization from cyphase's link 2011-10-23T22:37:50 my other things are multi source and reusing ants that have already collected a food item and will be usable in the future 2011-10-23T22:38:00 Guest42467, repeatedly check remaining time 2011-10-23T22:38:02 don't think that can be improved, tbh 2011-10-23T22:38:12 cross product optimization? :\ 2011-10-23T22:38:23 thanks lrh9 2011-10-23T22:38:39 Islacrusez: That is a script line which in linux/unix tells your system what shell to run the script with 2011-10-23T22:38:46 *** avers_ has joined #aichallenge 2011-10-23T22:38:52 antimatroid: http://theory.stanford.edu/~amitp/GameProgramming/Heuristics.html#breaking-ties 2011-10-23T22:39:40 is it just me or would that conflict with what he just said? 2011-10-23T22:39:49 hmm okay, i'll read through that later some time when i'm not meant to be doing an exam, thanks 2011-10-23T22:40:14 Islacrusez: he meant he doesn't think you can improve on the cross product optimization 2011-10-23T22:41:13 i'd just like to take this opportunity to make this public service announcement: If you're not using a VCS for your bot code, start now. 2011-10-23T22:41:17 now I'm confused... eh? 2011-10-23T22:41:28 what's a vcs? 2011-10-23T22:41:33 version control syste, 2011-10-23T22:41:34 version control system 2011-10-23T22:41:34 *** Saulzar has joined #aichallenge 2011-10-23T22:41:37 system* 2011-10-23T22:41:38 ah 2011-10-23T22:41:38 like git 2011-10-23T22:41:50 my vcs for ai is copy paste aha 2011-10-23T22:41:59 really, you should use github or bitbucket 2011-10-23T22:42:03 vcs + offsite backup 2011-10-23T22:42:03 lol 2011-10-23T22:42:04 :D 2011-10-23T22:42:10 also give us the link 2011-10-23T22:42:10 github has to be public for free 2011-10-23T22:42:16 orly? 2011-10-23T22:42:25 i guess just use bitbucket then :> 2011-10-23T22:42:39 you can use git on you local without publishing code to the internet 2011-10-23T22:42:41 rules for this, are code must be 100% your own? 2011-10-23T22:42:52 i'm also a complete noob with github 2011-10-23T22:42:55 antimatroid, do you have do_turn_26()? 2011-10-23T22:42:59 i can pull, add, commit and push :P 2011-10-23T22:43:11 and if something goes wrong i cry at amstan 2011-10-23T22:43:30 I'm a compelte tool with vcs systems.. I've successfully broken 4 different types of it 2011-10-23T22:43:43 cyphase: what? 2011-10-23T22:43:46 broken = have to take a trip to teh bank, to get backup tapes 2011-10-23T22:43:57 haha 2011-10-23T22:44:07 antimatroid, you said your vcs is copy-paste. i know i've done that before, rename old versions of functions as function_1() or whatever 2011-10-23T22:44:09 I can break any code, crash or corrupt any device 2011-10-23T22:44:31 i usually just named them as i got further into developing components of my bot 2011-10-23T22:44:40 like tree/decision/heuristic bot etc. 2011-10-23T22:44:56 that's like the reverse order of development :P 2011-10-23T22:46:09 *** otzi has quit IRC (Ping timeout: 260 seconds) 2011-10-23T22:47:31 @rankings 2011-10-23T22:47:33 antimatroid: Top 10 players: FlagCapper(76.5), GreenTea(76.2), MomoBot(75.3), sir_macelon(74.9), paratrechina(73.4), ThisIsNotABug(73.2), Ender(73.1), skyramp(72.6), ZumZoom(72.1), ChrisH(71.8) 2011-10-23T22:48:16 I just watched through some of their matches, they all have decent combat logic. :( 2011-10-23T22:48:27 Is HunterBot better than the Starting Kit? 2011-10-23T22:48:54 it performs better, but might lead you down a wrong road 2011-10-23T22:48:57 analyst74: bear in mind they've all been working on their bot since during beta 2011-10-23T22:49:18 Guest42467, HunterBot is better than the starter kit after going through the tutorial 2011-10-23T22:49:24 although ant hills were only added a couple of weeks ago 2011-10-23T22:49:26 on some maps at least 2011-10-23T22:49:47 if ant hills are the sole purpose, how were the only recently added? 2011-10-23T22:49:48 antimatroid, how were ants spawned before hills? at the site of the food? 2011-10-23T22:49:49 re added if you count my early prototype game :P 2011-10-23T22:50:11 used to just spawn where the food was, aim was to take out the other colonies 2011-10-23T22:50:20 and you scored points for spawninng ants and killing enemies 2011-10-23T22:50:29 it's way simpler to understand now and has a better objective imo 2011-10-23T22:50:53 i like to think of it being like future cop lapd precinct assault 2011-10-23T22:50:53 agree 2011-10-23T22:51:15 defend your base and invade the enemies bases 2011-10-23T22:51:15 ya sounds better, but I think you should gain the hill you take rather than raze it, but for games ending in due time razing is better 2011-10-23T22:51:32 that would be cool but makes games last longer possibly 2011-10-23T22:51:34 Its okay for me to upload the hunterbot to my account, right? 2011-10-23T22:51:38 and doesn't really lead to a natural conclusion 2011-10-23T22:51:50 it's be a lot more like planet wars if that was the case too, especially if we added in neutral hills 2011-10-23T22:51:59 Guest42467: sure, but why? :P 2011-10-23T22:52:21 Well, currently I have the base starting kit uploaded, so at least it'll be an improvement 2011-10-23T22:52:35 *** otzi has joined #aichallenge 2011-10-23T22:52:39 yeah but it takes a lot of the fun out if it's not a bot you wrote :P 2011-10-23T22:52:44 *** Saulzar has quit IRC (Ping timeout: 256 seconds) 2011-10-23T22:53:43 *** avdg has quit IRC (Quit: Leaving.) 2011-10-23T22:54:23 ok, this blows... in 7 hours I need to get up, in 12 hours I need to go and key... I am now able to run my bot which means in about 6 hours I'll get to sleep xD 2011-10-23T22:54:42 lol 2011-10-23T22:55:14 *** Guest42467 has quit IRC (Quit: Page closed) 2011-10-23T22:55:15 anyone know: are the rules as they currently exist, for sure the way they are going to be, for final? 2011-10-23T22:56:04 burny: it is undesirable for contestants and organisers that they do not change, but if the game is considered "broken" for some reason then we will change them 2011-10-23T22:56:23 *** avdg has joined #aichallenge 2011-10-23T22:56:34 however we have given parameters for most things in turn 0 such that those can be changed in such a situation, although that is undesirable too 2011-10-23T22:56:54 lol, it's desirable that they don't change :P 2011-10-23T22:57:11 *** garlicnation has joined #aichallenge 2011-10-23T22:57:33 Hardcode as little as reasonably possible, but I doubt the game will change from tiles to hexes for instance lol. 2011-10-23T22:57:37 ok, so they'll only change if game is broken.. and will change by minimal amount, in that case? 2011-10-23T22:57:45 yes 2011-10-23T22:57:59 more map generators will be made 2011-10-23T22:58:20 we need to work out a maximum dimensions at some point 2011-10-23T22:58:28 and actually use maps up to those dimensions in the final 2011-10-23T22:58:36 tetrehedrons 2011-10-23T22:58:43 that's right, 4D ANTS! 2011-10-23T22:58:47 err, 3D 2011-10-23T22:58:50 Haha. 2011-10-23T22:58:55 i hate vague rules if the final contest doesn't use the full generality that they allow 2011-10-23T22:58:57 arbitrarily make the game 3D two days before the final deadline xD 2011-10-23T22:59:01 lol 2011-10-23T22:59:07 ssshh 2011-10-23T22:59:15 *works on his 3d bot* 2011-10-23T22:59:16 i love vague rules otherwise 2011-10-23T22:59:16 4D would be cool 2011-10-23T22:59:19 so, for sure won't change for things like.. 2011-10-23T22:59:23 vague but unambiguous 2011-10-23T22:59:30 the map is made of hypercubes? 2011-10-23T22:59:31 all hills generate upon a food gather 2011-10-23T22:59:39 you could display the ants on a cube without making the physics 3d 2011-10-23T22:59:53 lol, physics 2011-10-23T22:59:58 the game is currently being played on the surface of a discrete torus, what more do you want? 2011-10-23T22:59:59 or, after gathering a food, the ant must go back to a hill, to spawn a new ant 2011-10-23T23:00:01 it does change the wraparound a bit I guess 2011-10-23T23:00:05 antimatroid: maps with a max dimension of 150 are already in use 2011-10-23T23:00:22 janzert: my point is we need to set a maximum and use maps up to that maximum at the end 2011-10-23T23:00:26 I also think that is hitting the limit of what can be visualized well 2011-10-23T23:01:11 exactly, i think we need to work out a feasible limit that people should expect their bots to run on, but then it sucks if we use only a subset of those rules in the final and people who made assumptions get an advantage 2011-10-23T23:01:53 just putting my vote in for 150 :) 2011-10-23T23:02:02 Hey janzert, did you see my pull request perchance? 2011-10-23T23:02:46 all hills, always have a path of atleast width 1, to all other hills? 2011-10-23T23:02:46 didn't get a chance to look at it yet no 2011-10-23T23:03:01 burny: yes 2011-10-23T23:03:15 okay, just bumping you is all :P 2011-10-23T23:03:17 i've tried to make it at least 3 with the maze maps, although i think i saw one the other day where that didn't happen 2011-10-23T23:03:25 otherwise blocking is possible 2011-10-23T23:03:37 the "what is near me" visibility stuff, will always allow you to see through ants/water/etc.? 2011-10-23T23:03:51 *** ashoka has quit IRC (Quit: Page closed) 2011-10-23T23:03:56 wtf is a tuple anyhow? 2011-10-23T23:04:12 Islacrusez, you know arrays? Sort of like that. 2011-10-23T23:04:29 AndrewBC: I believe one problem at the moment is the database field won't be big enough to hold it 2011-10-23T23:04:33 It's an immutable, ordered collection of objects. 2011-10-23T23:04:42 janzert: oh shoot I didn't think to check that 2011-10-23T23:04:50 but I'll try and look into it in detail a little later 2011-10-23T23:04:52 great, I was never good with arrays; but ok, it's basically a list of a description noted above 2011-10-23T23:04:57 thanks mate 2011-10-23T23:05:10 there are a few cases now that are overflowing the field 2011-10-23T23:06:34 *** freenity has joined #aichallenge 2011-10-23T23:06:49 oh, wut? I must've implemented the code wrong somewhere... it says I'm supposed to get the same moves? I don't... bah 2011-10-23T23:06:55 *** skyramp has quit IRC (Ping timeout: 265 seconds) 2011-10-23T23:07:49 *** cutoff has quit IRC (Ping timeout: 258 seconds) 2011-10-23T23:08:14 *** antong has quit IRC (Quit: Page closed) 2011-10-23T23:08:15 anti: do submissions have to be 100% own code, or can share non-logic portions with others(eg. updating board state) 2011-10-23T23:08:17 *** mleise has quit IRC (Ping timeout: 252 seconds) 2011-10-23T23:09:02 Islacrusez, your game gets a number to seed the random generator so that games can play out the same way over multiple runs. Useful for debugging purposes. You may fail to seed the generator, seed it differently, or use a generator that does not rely on a seed. 2011-10-23T23:09:08 Islacrusez: unfortunately I think the tutorial was made with viewradius2=55, you might want to try setting that at the command line of playgame.py with --viewradius2 55 2011-10-23T23:09:31 the default now is 77 2011-10-23T23:09:31 *** antsnub has joined #aichallenge 2011-10-23T23:09:36 I lol'd when I saw the tools scripts use 42 as the seed 2011-10-23T23:09:46 Ah. I thought he was talking about the same moves for multiple games. 2011-10-23T23:09:48 Sorry. 2011-10-23T23:09:51 well that's helpful... 2011-10-23T23:09:52 Disregard my comment. 2011-10-23T23:09:58 hey all, any idea how to get rid of this nasty g++ flag? warning: ignoring return value of 'int scanf(const char*, ...)', declared with attribute warn_unused_result 2011-10-23T23:09:59 the change in defaults, that is 2011-10-23T23:10:11 on server, of course :P 2011-10-23T23:10:32 *** Kingpin13 has quit IRC (Quit: quit) 2011-10-23T23:10:40 maybe if you wrap the scanf in an if statement to make sure you got data :P 2011-10-23T23:11:21 what fun is that :P 2011-10-23T23:11:30 AHA, thanks janzert 2011-10-23T23:11:34 but i guess that'll have to do for now 2011-10-23T23:11:42 *** mleise has joined #aichallenge 2011-10-23T23:12:03 tied up for a bit, but I'll try and remember to add a note to the tutorial later 2011-10-23T23:12:19 *** slayer has joined #aichallenge 2011-10-23T23:12:26 hopefully someone can redo it with the new viewradius eventually 2011-10-23T23:15:39 *** Larose has joined #aichallenge 2011-10-23T23:16:09 *** freenity has quit IRC (Remote host closed the connection) 2011-10-23T23:16:51 *** jtamer has joined #aichallenge 2011-10-23T23:18:18 C Devs: is it necessary to free malloc'ed stuff at the end phase? or will server take care of it 2011-10-23T23:19:04 * Islacrusez growls at the tutorial 2011-10-23T23:19:19 depends on the os.. but pretty much every os does it 2011-10-23T23:19:42 *** gdt has joined #aichallenge 2011-10-23T23:20:02 add this after this function, replace that function with this... and we're supposed to know where a function starts and ends how exactly >< 2011-10-23T23:20:04 I believe does did not 2011-10-23T23:20:16 add 'dos' in that sentance 2011-10-23T23:20:21 er.. does=dos 2011-10-23T23:21:02 *** gdt has quit IRC (Client Quit) 2011-10-23T23:21:12 noob friendly it is not, and that means I can do it in the morning, and by morning I mean after midday 2011-10-23T23:22:25 gnight folks 2011-10-23T23:22:40 *** antsnub has quit IRC (Quit: Page closed) 2011-10-23T23:22:52 *** lavalamp has joined #aichallenge 2011-10-23T23:22:58 Functions are defined using the keyword "def". 2011-10-23T23:23:31 yes but where do they end? 2011-10-23T23:24:27 whitespace! 2011-10-23T23:24:33 it's all about the indentation, yo 2011-10-23T23:24:38 this is the sort of question you read tutorials/references for 2011-10-23T23:24:41 Islacrusez, in Python whitespace is significant. Lines with the same indentation indicate they are in the same scope. Funtion definitions end when the next line has the same indentation as the def statement. 2011-10-23T23:24:58 analyst74: the tutorial tells you *nothing* 2011-10-23T23:25:12 http://docs.python.org/py3k/tutorial/index.html 2011-10-23T23:25:14 islacrusez: i think he meant a general python tutorial on the internet 2011-10-23T23:25:14 *** cutoff has joined #aichallenge 2011-10-23T23:25:26 islacrusez: you can use any programming language you want, pretty much. it doesn't have to be python. 2011-10-23T23:26:39 it would help if I actually knew a programming language; the closest I know is AS2 that I last used about 5 years ago and has now degenerated into vague pseudocode, and so python was recommended 2011-10-23T23:26:50 *** ztfw has joined #aichallenge 2011-10-23T23:27:56 huh, the server's go version really is set as release.r60.1 9753+ 2011-10-23T23:28:15 that's not good 2011-10-23T23:28:42 want to playyyy 2011-10-23T23:29:22 *** Nakyuant has joined #aichallenge 2011-10-23T23:29:37 and all I'm really saying is that rtfm is a bad response when someone's trying to follow the tutorial; a link to a more useful tut/reference on the other hand is a good answer (cookie for lrh9) 2011-10-23T23:29:47 and now I'm going to sleep because that damn thing is 14 chapters long 2011-10-23T23:30:04 * lrh9 has a cookie. 2011-10-23T23:30:13 python rulez 2011-10-23T23:30:26 lrh9 definitely deserved his cookie! 2011-10-23T23:31:03 good luck in learning python Islacrusez 2011-10-23T23:31:40 although I'm not using python 3 so actually that didn't help as much, but I'd imagine there's a 2.7 document somewhere on the same site 2011-10-23T23:31:53 Ah. Oops. 2011-10-23T23:32:03 *** djr__ has quit IRC (Quit: Leaving) 2011-10-23T23:32:11 I use 3 so it was a knee jerk reaction. 2011-10-23T23:33:35 *** yoden has quit IRC (Quit: Leaving.) 2011-10-23T23:33:50 found the 2.7 one 2011-10-23T23:33:58 *** pedrosorio has quit IRC (Quit: Page closed) 2011-10-23T23:34:44 you get to keep your cookie =P 2011-10-23T23:35:03 and now I'm going to sleep because otherwise I'll never survive work 2011-10-23T23:35:04 does the engine output any blank lines? 2011-10-23T23:35:13 cya folks, and thanks for the assistance 2011-10-23T23:35:14 oh man 2011-10-23T23:35:19 I'm number dyslexic 2011-10-23T23:35:28 I do love Python. I tried C++ years ago, but I never good get into it. I have nothing against C++. In fact I want to really learn it. But I think the technical complexity distracted me from what I believe to be the important foundations of programming: design, design, and more design. 2011-10-23T23:35:37 the site's go is 9753, I accidentally checked out 9573 and compiled that 2011-10-23T23:35:41 s/good/could 2011-10-23T23:36:13 *** Islacrusez has left #aichallenge 2011-10-23T23:36:20 I wish there was a programing language, that was based on design 2011-10-23T23:36:49 "make shiny button" 2011-10-23T23:36:54 "round its corners" 2011-10-23T23:37:02 GRAET SUCCESS 2011-10-23T23:37:11 I wonder if there's any company that would want to put a few million$, into a set of design+dev tools 2011-10-23T23:37:57 *** Nakyuant has quit IRC (Quit: Page closed) 2011-10-23T23:38:27 I don't really use modeling tools, but as I understand it they are supposed to help. E.g. UML. I don't know how effective they are though. 2011-10-23T23:38:55 uml is only part of the design 2011-10-23T23:39:30 and, because uml isn't integrated with the implementation.. means you can't design and dev simultaneously 2011-10-23T23:39:58 meh, UML is too much 2011-10-23T23:40:07 uml is too little by my books 2011-10-23T23:40:25 the problem is obvious from its very first initial 2011-10-23T23:40:45 *** slayer has quit IRC (Quit: Page closed) 2011-10-23T23:41:19 *** teryret has joined #aichallenge 2011-10-23T23:41:48 so the posted rules say "no threading", does that mean no go routines either? 2011-10-23T23:41:53 I appreciate what the creators are trying to do, but when someone has to explain to someone what a symbol in the model means because they've never seen it before because it's too broad a modeling language, then you haven't really saved any effort by putting it in the model than just explaining with a label and a circle anyway 2011-10-23T23:42:04 plus, the books they write are as boring as fuck 2011-10-23T23:42:16 teryret: no goroutines are fine 2011-10-23T23:42:24 they're just multiplexed on top of 1 thread 2011-10-23T23:42:29 excellent, thanks! 2011-10-23T23:42:32 so you still get concurrency, but no parallelism 2011-10-23T23:42:41 yarr 2011-10-23T23:42:45 My design process consists of thinking about the problem really hard and writing/rewriting implementations until I get one I'm satisfied with. :) 2011-10-23T23:42:49 but more importantly, I finally get an excuse to go learn go 2011-10-23T23:42:50 from my debug output: 2011/10/24 03:26:40 GOMAXPROCS: %!s(int=1) 2011-10-23T23:42:50 2011/10/24 03:26:40 Go Version: release.r60.1 9753+ 2011-10-23T23:43:19 I haven't yet felt brave enough to attempt setting that to a higher number ;x 2011-10-23T23:43:38 I'm afraid I don't know go well enough to appreciate what you're showing me 2011-10-23T23:44:08 lol, that's the runtime value that determines the number of processors (virtual or otherwise) that can be used by the scheduler 2011-10-23T23:44:09 the advantage to the uml that has to be explained.. is you only have to explain it once 2011-10-23T23:44:17 burny: nonsense :P 2011-10-23T23:44:49 yeah, I don't have the context, but I do know that I have to go re-learn UML every time I have to use it 2011-10-23T23:44:53 it's a language, it's likely that something that needs to be explained will need to be explained again before it's retained, just like any other language 2011-10-23T23:44:55 I want the design to be integrated with the implimentation though 2011-10-23T23:45:22 why not do it reactively then, a tool that allows you to visualize what you've written? 2011-10-23T23:45:27 There are attempts to use UML for automatic code generation. 2011-10-23T23:45:44 ^ more "universal" 2011-10-23T23:45:46 *** ajf|offline has quit IRC (Ping timeout: 258 seconds) 2011-10-23T23:45:50 eg. part of the design details things like.. what other data it is dependant on, if it is read-only, or if it modifies the data, if other instances of that class, can interact with the same isntances of some dependancy, etc. 2011-10-23T23:45:58 let's do everything so we can be shitty at most of it 2011-10-23T23:46:18 *** ajf|offline has joined #aichallenge 2011-10-23T23:46:21 uml -> code, isn't integration(or really of much use) 2011-10-23T23:46:23 I'm sorry, Andrew, how did we get onto Java? 2011-10-23T23:46:23 *** twymer has quit IRC (Ping timeout: 248 seconds) 2011-10-23T23:46:35 I'm talking about UML 2011-10-23T23:46:44 *** ajf|offline is now known as Guest21952 2011-10-23T23:46:48 I know ;-) applies to both 2011-10-23T23:46:55 oh I see what you did thar 2011-10-23T23:47:08 It's probably not surprising that I dislike java too 2011-10-23T23:47:14 *** Tommygun has joined #aichallenge 2011-10-23T23:47:17 I also dislike C++ 2011-10-23T23:47:18 Hello 2011-10-23T23:47:20 I gathered by the fact that you did it in go 2011-10-23T23:47:25 :P 2011-10-23T23:47:30 *** Boxx has joined #aichallenge 2011-10-23T23:47:33 How is everyone tonight? 2011-10-23T23:47:34 I ALMOST like c++ 2011-10-23T23:47:36 yeah, I'm not a super huge fan of c++ either 2011-10-23T23:47:44 if you have a good idea about the software that will get produced by a tool, then you can create a very easy to use interface. Something that you don't even need to be a programmer to use. 2011-10-23T23:47:50 I slightly like c# 2011-10-23T23:47:58 language wars is not my thing though, I started with PHP so I don't berate people for their choices 2011-10-23T23:47:58 ditto 2011-10-23T23:48:15 I'd like C# a lot more if I didn't dislike the problems it is super good at solving 2011-10-23T23:48:24 If you have a deep understanding of the generated code, all you can do is offer the programmer the possibility of doing it. But he'll need ot think by himself. 2011-10-23T23:49:21 i am a fan of simple, clear specifications. for that reason, i am not a fan of uml. 2011-10-23T23:49:29 here here 2011-10-23T23:49:41 i'm half joking 2011-10-23T23:49:54 well, it's definitely a specification 2011-10-23T23:49:59 lol 2011-10-23T23:50:03 Agh. Bot has a bug. 2011-10-23T23:50:21 I'm not fan of specifications, but I like easy to read, of course :P 2011-10-23T23:51:41 a complete design, is sufficient to allow a computer to do a complete implementation 2011-10-23T23:52:04 i think i should not have used such an overloaded word in my claim 2011-10-23T23:52:20 the thing is that reading and writing specifications is a really boring task 2011-10-23T23:52:24 so currently nothing besides programming languages, allow you to do complete designs 2011-10-23T23:52:36 but it's important for programmer interaction, of course 2011-10-23T23:52:38 and we should keep it that way 2011-10-23T23:52:45 you can write programs to enforce specifications 2011-10-23T23:52:50 burny, WYSIWYG? 2011-10-23T23:53:17 what better way to encode a specification than into a rigid unforgiving program that must be used? 2011-10-23T23:53:35 if I lived in the world of Tron, I'd be a harsh dictator 2011-10-23T23:53:40 currently, I do all my design in c++ 2011-10-23T23:53:48 hehehe 2011-10-23T23:53:51 it's sort of annoying for huge projects 2011-10-23T23:54:05 you joke burny, but at my last job we did our web templates in c++, I shit you not 2011-10-23T23:54:12 lolwut 2011-10-23T23:54:12 formal specifications don't necessarily give enough detail to make the resulting program efficient 2011-10-23T23:54:45 yeah, when you really really _really_ need your page to be as fast as humanly possible, you have to use native code 2011-10-23T23:54:48 I do all software design in a whiteboard, then we draw a bunch of things on paper and whatnot 2011-10-23T23:54:55 and we needed it to be that fast 2011-10-23T23:55:03 then we throw it all away and write something completely different 2011-10-23T23:55:16 sometimes for the better, sometimes for the worse 2011-10-23T23:55:31 jmc: a complete design would be sufficient to allow efficient generation 2011-10-23T23:55:43 burny: then that's a program, not a "design" 2011-10-23T23:56:06 the line is so blurry it probably doesn't exist 2011-10-23T23:56:23 well, technically.. code is design 2011-10-23T23:56:27 burny: to be concrete, i'm talking about specification in the form of logical propositions about the program's behavior at a high level. 2011-10-23T23:56:33 code is design implementation 2011-10-23T23:56:56 to me, anything lower level than that is leaving the world of specification and moving into the world of development 2011-10-23T23:57:09 what you're really saying is you design in your head and then implement it 2011-10-23T23:57:16 which is how I roll, by the way 2011-10-23T23:59:16 for example, i can perfectly specify what set union means, but there are still many different ways to implement it, each with different performance characteristics 2011-10-23T23:59:31 that doesn't mean it's underspecified, though 2011-10-23T23:59:36 *** Tommygun has quit IRC (Ping timeout: 265 seconds) 2011-10-23T23:59:51 ya.. and a full design, would allow the computer to determine which implementation is best